KR102342314B1 - Method, server and computer program for providing answers to inquiry data based on quality data of pharmaceuticals - Google Patents

Abstract

Provided are a method, a server and a computer program for providing responses to query data based on the quality data of pharmaceutical products. A method of providing a response to query data based on quality data of medicines according to various embodiments of the present invention is a method performed by a computing device, the method comprising: obtaining query data; metadataizing the query data to obtain query metadata; selecting one or more similar data sets by performing a search on the drug quality database based on the query metadata; performing data grouping through classification of a plurality of quality data corresponding to the one or more similar data sets and the query data; and providing drug analysis information corresponding to the query data based on the data grouping result.

Description

Method, server and computer program to provide answers to inquiry data based on quality data of pharmaceuticals

Various embodiments of the present invention relate to a method, a server, and a computer program for providing a response to query data based on quality data of pharmaceuticals.

Pharmaceuticals generally refer to any substance used to alter or examine physiological systems or disease states for the benefit of users. Pharmaceuticals may include synthetic drugs and biologics. Biopharmaceuticals (i.e., biologics) are pharmaceuticals manufactured using raw materials or materials derived from humans or other living organisms, and refer to pharmaceuticals that require special attention in terms of health and hygiene. Includes therapeutics, gene therapy products, and other products approved by the Minister of Food and Drug Safety.

These drugs can be managed through the Pharmaceutical Quality Management System (QMS), which seeks to maintain and improve the quality by finding the cause of the occurrence of defective products and removing them by using statistical methods for production management. The pharmaceutical quality management system refers to the totality of pharmaceutical and bio business processes that focus on achieving quality policies and quality objectives to meet user requirements, stability and efficacy of pharmaceuticals.

Recently, the use of electronic medicine quality management system (eQMS), which computerized the pharmaceutical quality management system, is increasing. eQMS refers to software that maximizes the efficiency of work processes by digitalizing the document-based pharmaceutical quality management system. For example, the document-based drug quality management system has inefficiencies in various aspects, such as the possibility of loss of documents, the possibility of omission of item descriptions, or difficulty in accessing documents, and eQMS handles these problems efficiently make it possible

On the other hand, in the traditional process validation/control strategy, quality and process control strategies were established with a limited number of validation batches. In order to fundamentally solve this problem, global pharmaceutical regulatory agencies have come to actively recommend the introduction of Quality by Design (QbD).

Accordingly, through eQMS, in accordance with the requirements of global pharmaceutical regulatory agencies, technological developments are being made to predict various issues that occur in the R&D and production process of pharmaceuticals from a large amount of electronic quality data for overall pharmaceuticals. .

Republic of Korea Patent 10-2274363

The problem to be solved by the present invention has been devised in response to the above-mentioned background technology, and it is possible to construct big data based on the quality data of medicines, and to provide answers to questions that arise in the process of drug R&D and production. To provide an artificial intelligence model.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the following description.

Disclosed is a method for providing a response to query data based on quality data of pharmaceuticals according to an embodiment of the present invention for solving the above-described problems. The method includes: obtaining query data; performing a search on a drug quality database based on the query data to select one or more similar data sets; a plurality of quality data corresponding to the one or more similar data sets; and It may include performing data grouping through classification of the query data, and providing drug analysis information corresponding to the query data based on the data grouping result.

In an alternative embodiment, further comprising the step of building the drug quality database, the step of building the drug quality database, a plurality of quality data (OQD, Overall Quality Data) corresponding to each of a plurality of drugs to obtain and grouping each of the plurality of quality data into one or more data sets based on a Critical Quality Profile (CQP) corresponding to each of the plurality of quality data.

In an alternative embodiment, the quality data includes general information of the drug, profile data related to production and quality definition, and empirical data corresponding to the profile data, wherein the main quality profile includes characteristics and properties of the drug. Information on key factors that can be determined, used for the search, and may be configured through at least a part of the quality data.

In an alternative embodiment, the step of building the drug quality database includes learning about the correlation between the elements constituting each of the plurality of quality data through the association rule analysis algorithm to determine the correlation between the respective elements. It may include the steps of generating a correlation analysis model to be derived, and performing metadata on the plurality of quality data based on the correlation between the respective elements to construct the drug quality database.

In an alternative embodiment, the selecting of the one or more similar data sets may include selecting one or more similar data sets having a similarity equal to or greater than a threshold similarity score to a key quality profile corresponding to the query data in the drug quality database. may include

In an alternative embodiment, the threshold similarity score is calculated based on a similarity score between key quality profiles corresponding to each of the one or more data sets, and based on one or more similarity scores generated corresponding to each of the data set pairs. can be characterized as

In an alternative embodiment, performing data grouping through classification on the plurality of quality data and the query data corresponding to the one or more similar data sets includes each of the plurality of quality data corresponding to the one or more similar data sets. It may include identifying one or more elements constituting the , and performing data grouping by classifying the plurality of quality data and the query data into one or more data groups, respectively, based on the one or more elements.

In an alternative embodiment, the providing of the drug analysis information may include deriving a correlation between a data group in which the query data is classified among the one or more data groups and each of the remaining data groups by utilizing the correlation analysis model. and providing drug analysis information corresponding to the query data based on the correlation between the respective data groups.

According to another embodiment of the present disclosure, a computing device for performing a method of providing a response to query data based on quality data of a drug is disclosed. The computing device includes a storage unit for storing one or more instructions and a processor for executing one or more instructions stored in the storage unit, wherein the processor executes the one or more instructions, thereby querying based on the quality data of the above-mentioned medicines. You can perform a method that provides a response to the data.

According to another embodiment of the present invention, a computer program stored in a computer-readable recording medium is disclosed. The computer program may be combined with a computer that is hardware to perform a method of providing a response to the query data based on the quality data of the above-described medicines.

Other specific details of the invention are included in the detailed description and drawings.

According to various embodiments of the present invention, it is possible to provide an artificial intelligence model that can provide answers to questions that occur in the process of drug R&D and production.

Effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is an exemplary diagram schematically illustrating a system for providing a response to inquiry data based on drug quality data related to an embodiment of the present invention.
2 is a block diagram of a server that provides a response to query data based on drug quality data related to an embodiment of the present invention.
3 is a flowchart exemplarily illustrating a process of building a database by acquiring a plurality of quality data related to an embodiment of the present invention.
4 is a flowchart exemplarily illustrating a process of providing a response corresponding to the corresponding query data based on the query data related to an embodiment of the present invention.
5 is an exemplary diagram illustrating a process of generating one or more data groups based on one or more similar data sets related to an embodiment of the present invention.
6 is a flowchart exemplarily illustrating a method of providing a response to inquiry data based on drug quality data related to an embodiment of the present invention.
7 is a schematic diagram illustrating one or more network functions related to an embodiment of the present invention.

Various embodiments are now described with reference to the drawings. In this specification, various descriptions are presented to provide an understanding of the present disclosure. However, it is apparent that these embodiments may be practiced without these specific descriptions.

The terms “component,” “module,” “system,” and the like, as used herein, refer to a computer-related entity, hardware, firmware, software, a combination of software and hardware, or execution of software. For example, a component can be, but is not limited to being, a process running on a processor, a processor, an object, a thread of execution, a program, and/or a computer. For example, both an application running on a computing device and the computing device may be a component. One or more components may reside within a processor and/or thread of execution. A component may be localized within one computer. A component may be distributed between two or more computers. In addition, these components can execute from various computer readable media having various data structures stored therein. Components may communicate via a network such as the Internet with another system, for example, via a signal having one or more data packets (eg, data and/or signals from one component interacting with another component in a local system, distributed system, etc.) may communicate via local and/or remote processes depending on the data being transmitted).

In addition, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or.” That is, unless otherwise specified or clear from context, "X employs A or B" is intended to mean one of the natural implicit substitutions. That is, X employs A; X employs B; or when X employs both A and B, "X employs A or B" may apply to either of these cases. It should also be understood that the term “and/or” as used herein refers to and includes all possible combinations of one or more of the listed related items.

Also, the terms "comprises" and/or "comprising" should be understood to mean that the feature and/or element in question is present. However, it should be understood that the terms "comprises" and/or "comprising" do not exclude the presence or addition of one or more other features, elements and/or groups thereof. Also, unless otherwise specified or unless it is clear from context to refer to a singular form, the singular in the specification and claims should generally be construed to mean “one or more”.

Those skilled in the art will further appreciate that the various illustrative logical blocks, configurations, modules, circuits, means, logics, and algorithm steps described in connection with the embodiments disclosed herein may be implemented in electronic hardware, computer software, or combinations of both. It should be recognized that they can be implemented with To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, configurations, means, logics, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application. However, such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.

Descriptions of the presented embodiments are provided to enable those of ordinary skill in the art to use or practice the present disclosure. Various modifications to these embodiments will be apparent to those skilled in the art of the present disclosure. The generic principles defined herein may be applied to other embodiments without departing from the scope of the present disclosure. Thus, the present disclosure is not limited to the embodiments presented herein. This disclosure is to be interpreted in the widest scope consistent with the principles and novel features presented herein.

In this specification, a computer refers to all types of hardware devices including at least one processor, and may be understood as encompassing software configurations operating in the corresponding hardware device according to embodiments. For example, a computer may be understood to include, but is not limited to, smart phones, tablet PCs, desktops, notebooks, and user clients and applications running on each device.

Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

Each step described in this specification is described as being performed by a computer, but the subject of each step is not limited thereto, and at least a portion of each step may be performed in different devices according to embodiments.

1 is an exemplary diagram schematically illustrating a system for providing a response to inquiry data based on drug quality data related to an embodiment of the present invention. As shown in FIG. 1 , a system 10 according to embodiments of the present invention may include a server 100 , a client 200 , and a network. The components illustrated in FIG. 1 are exemplary, and additional components may be present or some of the components illustrated in FIG. 1 may be omitted. The server 100 and the client 200 according to the embodiments of the present invention may mutually transmit and receive data for the system according to the embodiments of the present invention through a network.

Networks according to embodiments of the present invention include Public Switched Telephone Network (PSTN), x Digital Subscriber Line (xDSL), Rate Adaptive DSL (RADSL), Multi Rate DSL (MDSL), Very High Speed DSL (VDSL). ), a variety of wired communication systems such as Universal Asymmetric DSL (UADSL), High Bit Rate DSL (HDSL), and Local Area Network (LAN) can be used.

In addition, the networks presented herein include Code Division Multi Access (CDMA), Time Division Multi Access (TDMA), Frequency Division Multi Access (FDMA), Orthogonal Frequency Division Multi Access (OFDMA), Single Carrier-FDMA (SC-FDMA) and Various wireless communication systems may be used, such as other systems.

The network according to the embodiments of the present invention can be configured regardless of its communication mode, such as wired and wireless, and is composed of various communication networks such as a personal area network (PAN) and a wide area network (WAN). can be In addition, the network may be a well-known World Wide Web (WWW), and may use a wireless transmission technology used for short-range communication, such as infrared (IrDA) or Bluetooth (Bluetooth). The techniques described herein may be used in the networks mentioned above, as well as in other networks.

The client 200 according to an embodiment of the present invention may mean any type of node(s) in a system having a mechanism for communication with the server 100 . For example, such a client 200 may include a PC, a laptop computer, a workstation, a terminal, and/or any electronic device having network connectivity. In addition, the client may include any server implemented by at least one of an agent, an application programming interface (API), and a plug-in. In addition, the client 200 may include an application source and/or a client application. According to an embodiment of the present disclosure, operations to be described later of the server 100 may be performed according to a query (eg, query data) issued from the client 200 .

According to an embodiment of the present invention, the server 100 may provide a response corresponding to the query data received from the client. In the present invention, the query data may be data related to a query that occurs during drug R&D and production process. For example, the query data may be information about the quality data of a new drug under review that is not defined or stored in the existing database. As a specific example, the query data may relate to overall conditions (eg, set temperature, humidity, experiment time, etc. related to the experimental environment) for quality verification experiments related to a specific drug. For another example, the query data may include at least some of the quality characteristics of a specific drug (eg, an aggregation corresponding to a product-related impurity that is one of the quality characteristics of a drug), or a cut product ( truncated form) or a glycosylation pattern corresponding to structural properties). The detailed description of the above-described query data is only an example, and the present invention is not limited thereto.

That is, the server 100 may provide a response corresponding to the query data when receiving the query data related to the query that may be generated during the research and development and production process of the drug from the client. Here, the response corresponding to the inquiry data may be medicine analysis information corresponding to the inquiry data. For example, when the first query data is data related to a query related to product and process design, drug analysis information corresponding to the first query data is used to generate predicted result data and optimal result data according to the query data. It may include input data and adjustment data suggesting adjustment of the relationship between each element. For another example, if the second query data is data related to a query related to quality trend analysis for a trend element, drug analysis information corresponding to the second query data may be out-of-specification or trend It may include warning signs about the possibility of out-of-trends and information on follow-up actions. For another example, if the third inquiry data is data related to the inquiry about the stage of the production process of a drug, the drug analysis information corresponding to the third inquiry data is related to the risk factor with high risk among the major factors for each process stage. It may include information on related information and mitigation plans for mitigating high risks. The description of the aforementioned various query data and drug analysis information corresponding to each query data is only an example, and the present invention is not limited thereto. That is, the server 100 of the present invention may present various responses as described above in response to inquiries generated in the process of R&D and production of pharmaceuticals.

According to one embodiment of the present invention, server 100 may include any type of computer system or computer device, such as, for example, a microprocessor, a mainframe computer, a digital single processor, a portable device and a device controller, and the like. . Although not shown in FIG. 1 , the server 100 may include a database management system (DBMS). Additionally, server 100 may be used interchangeably with a device for executing queries. DBMS is a program for allowing the server 100 to perform operations such as parsing a query, searching for, inserting, modifying and/or deleting necessary data, and the processor in the storage unit 120 of the database server 100 (130) can be implemented.

The server 100 may include, but is not limited to, a device including a processor 130 and a storage unit 120 for executing and storing instructions as any type of database. That is, the server 100 may include software, firmware, hardware, or a combination thereof. The software may include application(s) for creating, deleting and modifying database tables, schemas, indexes and/or data. Server 100 may receive transactions from clients or other computing devices, exemplary transactions retrieving, inserting, modifying, deleting, and/or managing records, such as data, tables and/or indexes, and the like, at server 100 . may include doing

In addition, the server 100 may store information about a plurality of medicines. For example, the server 100 may be a server that acquires and stores a plurality of quality data related to each of a plurality of medicines and a main quality profile corresponding to each quality data. Information stored in the server 100 may be utilized as learning data, verification data, and test data for learning the neural network in the present invention. That is, the server 100 may store information about a data set for training the neural network model of the present invention.

The server 100 of the present invention may generate a correlation analysis model of the present invention by learning a neural network model through a plurality of quality data and a main quality profile corresponding to each quality data. The correlation analysis model may be a neural network model for analyzing correlations between a plurality of elements included in a plurality of quality data.

In addition, although only one server 100 is shown in FIG. 1 , it is in the field of the application that more servers may also be included in the scope of the present invention and that the server 100 may include additional components. It will be clear to those of ordinary skill in the art. That is, the server 100 may be composed of a plurality of computing devices. In other words, a set of a plurality of nodes may constitute the server 100 .

According to an embodiment of the present invention, the server 100 may be a server that provides a cloud computing service. More specifically, the server 100 is a type of Internet-based computing, and may be a server that provides a cloud computing service that processes information not with a user's computer but with another computer connected to the Internet. The cloud computing service may be a service that stores data on the Internet and allows the user to use it anytime and anywhere through Internet access without installing necessary data or programs on his/her computer. Easy to share and deliver with a click. In addition, cloud computing service not only stores data on a server on the Internet, but also allows users to perform desired tasks using the functions of applications provided on the web without installing a separate program, and multiple people can simultaneously view documents. It may be a service that allows you to work while sharing. In addition, the cloud computing service may be implemented in the form of at least one of Infrastructure as a Service (IaaS), Platform as a Service (PaaS), Software as a Service (SaaS), a virtual machine-based cloud server, and a container-based cloud server. . That is, the server 100 of the present invention may be implemented in the form of at least one of the above-described cloud computing services. The detailed description of the above-described cloud computing service is merely an example, and may include any platform for constructing the cloud computing environment of the present invention.

In the present invention, the learning method, learning process, and method of acquiring a plurality of quality data, building a drug quality database based on the acquired quality data, and when receiving query data, the drug quality database and the neural network model are utilized for the neural network in the present invention. Accordingly, a detailed description of a method of providing drug analysis information corresponding to the corresponding query data will be described later with reference to FIG. 2 below.

2 is a block diagram of a server for providing a method for providing a response to inquiry data based on quality data of medicines related to an embodiment of the present invention.

As shown in FIG. 2 , the server 100 may include a network unit 110 , a storage unit 120 , and a processor 130 . Components included in the above-described server 100 are exemplary and the scope of the present invention is not limited to the above-described components. That is, additional components may be included or some of the above-described components may be omitted depending on implementation aspects of the embodiments of the present disclosure.

According to an embodiment of the present invention, the server 100 may include a network unit 110 for transmitting and receiving data to and from the client 200 . The network unit 110 may transmit and receive data for performing the method for providing a response to query data according to an embodiment of the present invention and a training data set for training a neural network model with other computing devices, servers, etc. have. That is, the network unit 110 may provide a communication function between the server 100 and the client. For example, the network unit 110 may receive query data related to a specific medicine from the client 200 . As another example, the network unit 110 may receive metadata indexed in response to the query data of the present invention from the cloud server. Additionally, the network unit 110 may allow information transfer between the server 100 and the client 200 by calling a procedure to the server 100 .

Network unit 110 according to an embodiment of the present invention is a public switched telephone network (PSTN), xDSL (x Digital Subscriber Line), RADSL (Rate Adaptive DSL), MDSL (Multi Rate DSL), VDSL ( A variety of wired communication systems such as Very High Speed DSL), Universal Asymmetric DSL (UADSL), High Bit Rate DSL (HDSL), and Local Area Network (LAN) can be used.

In addition, the network unit 110 presented herein is CDMA (Code Division Multi Access), TDMA (Time Division Multi Access), FDMA (Frequency Division Multi Access), OFDMA (Orthogonal Frequency Division Multi Access), SC-FDMA ( A variety of wireless communication systems can be used, such as Single Carrier-FDMA) and other systems.

In the present invention, the network unit 110 may be configured regardless of its communication mode, such as wired and wireless, and may be composed of various communication networks such as a short-range network (PAN: Personal Area Network) and a local area network (WAN: Wide Area Network). can In addition, the network may be a well-known World Wide Web (WWW), and may use a wireless transmission technology used for short-range communication, such as infrared (IrDA) or Bluetooth (Bluetooth). The techniques described herein may be used in the networks mentioned above, as well as in other networks.

According to an embodiment of the present invention, the storage unit 120 may include a permanent storage medium and a memory.

Persistent storage media may include any data, such as, for example, magnetic disks, optical disks and magneto-optical storage devices, as well as storage devices based on flash memory and/or battery-backed memory. It may mean a non-volatile storage medium that can continuously perform Such a persistent storage medium may communicate with the processor 130 and the memory of the server 100 through various communication means. In an additional embodiment, such a persistent storage medium may be located outside the server 100 to be able to communicate with the server 100 .

Memory is the primary storage device directly accessed by the processor, such as random access memory (RAM), such as dynamic random access memory (DRAM), static random access memory (SRAM), etc. It may mean a volatile storage device in which stored information is momentarily erased, but is not limited thereto. Such a memory may be operated by the processor 130 . The memory may temporarily store a data table including data values. The data table may include data values, and in an embodiment of the present disclosure, the data values of the data table may be recorded from a memory to a persistent storage medium. In a further aspect, the memory includes a buffer cache, wherein data may be stored in data blocks of the buffer cache. Data stored in the buffer cache may be written to the persistent storage medium by a background process.

According to an embodiment of the present invention, the processor 130 may be configured with one or more cores, and may include a central processing unit (CPU) and a general purpose graphics processing unit (GPGPU) of a computing device. , data analysis such as a tensor processing unit (TPU), and a processor for deep learning.

The processor 130 may read the computer program stored in the storage unit 120 to perform data processing for deep learning according to an embodiment of the present invention. According to an embodiment of the present invention, the processor 130 may perform an operation for learning the neural network. The processor 130 for learning of the neural network, such as processing input data for learning in deep learning (DL), extracting features from input data, calculating an error, updating the weight of the neural network using backpropagation calculations can be performed.

Also, in the processor 130, at least one of a CPU, a GPGPU, and a TPU may process learning of a network function. For example, the CPU and the GPGPU can process learning of a network function and data classification using the network function. Also, in an embodiment of the present invention, learning of a network function and data classification using the network function may be processed by using the processors of a plurality of computing devices together. In addition, the computer program executed in the computing device according to an embodiment of the present invention may be a CPU, GPGPU or TPU executable program.

In the present specification, a network function may be used interchangeably with an artificial neural network and a neuron network. In the present specification, the network function may include one or more neural networks, and in this case, the output of the network function may be an ensemble of the outputs of the one or more neural networks.

The processor 130 may read a computer program stored in the storage unit 120 to provide a correlation analysis model according to an embodiment of the present invention. According to an embodiment of the present invention, the processor 130 may generate drug analysis information corresponding to the query data. According to an embodiment of the present invention, the processor 130 may perform a calculation for learning the correlation analysis model.

According to an embodiment of the present invention, the processor 130 may typically process the overall operation of the server 100 . The processor 130 processes signals, data, information, etc. input or output through the above-described components or runs an application program stored in the storage unit 120 to provide appropriate information or functions to a user or user terminal, or can be processed

According to an embodiment of the present invention, the processor 130 may build a drug quality database. The processor 130 may acquire a plurality of quality data (OQD, Overall Quality Data) corresponding to each of a plurality of medicines, and may build a medicine quality database based on the acquired plurality of quality data. Quality data corresponding to pharmaceuticals means all data for R&D, production and quality control from basic information about pharmaceuticals. It may include empirical data corresponding to . Profile data is, for example, project name/code, product name/code, product type/form, therapeutic area, indications, trademark/brand name, dosage form, route of administration Information on route of administration, excipients, shelf-life, development phase, manufacturing site, batch release site and target market, etc. Product quality objective consisting of general information including general information and product understanding information including information on product quality attributes, critical quality attributes, and acceptable ranges for quality attributes (Quality Target Product Profile) information may be included. In addition, the profile data is, for example, Process Flow/Process Unit Operations, Input Materials per Operation, Manufacturing Scale, Key Process Attributes , Process Parameters, Critical Process Parameters, Material Attributes, Critical Material Attributes, Process targets for quality attributes, Proven acceptable ranges) and information about the design space (Design Space), etc. may include information about the process understanding (Process Understanding). In addition, the profile data may be, for example, Input Materials Controls, Procedural Controls, Process Parameter controls, In-process Testing, Specifications, Information on Control Strategy Elements configured through information on Characterization & Comparability Testing and Process Monitoring may be included. The detailed description of the above-described profile data is only an example, and the present invention is not limited thereto.

The empirical data is data related to the actual observation environment or observation values of each individual batch of actual R&D or production corresponding to the profile data, for example, for each batch corresponding to the product quality target (Quality Target Product Profile). Production that can include actual observed values, and in addition to information on Material Controls, Equipment Controls, Process Controls, Hygiene and Cleaning for each batch (Manufacture) related information, facility management information, material and reagent reference data, substances and Quality control information and deviations, corrective and preventive actions, change control reports, and self-inspection data organized through information on reagent equivalence test data, operator records, labeling records, weighing records, instrument operation results, instrument calibration records, instrument set values, etc. , environmental monitoring checklist, environmental monitoring results, validation data, all causes of deviation from standards, results of deviations from standards, worker training records, production-related list evaluation records, etc. may include product warranty related information. The detailed description of the above-described empirical data is merely an example, and the present invention is not limited thereto. A specific method for the processor 130 to build a pharmaceutical quality database based on a plurality of quality data composed of profile data and empirical data as described above will be described below with reference to FIG. 3 .

3 is a flowchart exemplarily illustrating a process of building a database by acquiring a plurality of quality data related to an embodiment of the present invention.

The processor 130 may acquire a plurality of quality data (S110). Specifically, the processor 130 may acquire a plurality of quality data corresponding to each of a plurality of medicines. According to one embodiment, the processor 130 may convert the electronic form corresponding to the electronic pharmaceutical quality management system (eQMS) through a document scan of the quality document corresponding to the medicine. That is, the processor 130 may scan the quality documents corresponding to each medicine to form an image file. In addition, the processor 130 may perform optical character recognition (OCR) on the quality document converted into the electronic format to convert it to an electronic format corresponding to the eQMS. In this case, as the conversion to the electronic form is automated through OCR, the user's convenience may be increased.

According to another embodiment, the processor 130 may acquire a plurality of quality data by receiving an input corresponding to the document written in the form through the eQMS. Specifically, the eQMS may include an input window for allowing an electronic form input from a user. The user may input an input value corresponding to the format character in each item of the corresponding input window, and the processor 130 may obtain quality data by electronically documenting the input value based on the input value. For example, when the conversion to the electronic form is performed based on a user's input, a separate document scanning device or an optical character reading device is not necessarily required, and thus computing power can be reduced.

In addition, the processor 130 may perform grouping on a plurality of quality data ( S120 ). Specifically, the processor 130 may group each of the plurality of quality data into one or more data sets based on a Critical Quality Profile (CQP) corresponding to each of the plurality of quality data. The key quality profile may be information on key factors that can determine the characteristics and properties of a drug product. Such a main quality profile may be configured through at least a part of the quality data. For example, the main quality profile may be configured through at least some of a plurality of elements related to the profile data. For example, the main quality profile may be configured through information on general information and process understanding among profile data corresponding to a specific drug. As a more specific example, key quality profiles include product name/code, product type/form, indications, route of administration, process flow/process unit operations, and manufacturing scale ( Manufacturing Scale), etc. may be configured by including elements related to it. The key quality profile may be utilized for retrieval (or indexing) of quality data. The main quality profile may be information composed of main elements related to the characteristics and properties of a drug among a plurality of elements (or items) included in the quality data.

In one embodiment, a plurality of elements constituting the main quality profile (ie, main elements that can determine the characteristics and properties of a drug) may be arbitrarily added or changed by a user involved in the development and production of a drug. . According to an additional embodiment, a plurality of elements constituting the main quality profile may be determined by a neural network model trained based on deep learning. Specifically, the learned neural network model may be generated through learning based on various quality data. Such a neural network model may analyze the correlation between a plurality of elements included in specific quality data, and generate a major quality profile corresponding to the corresponding quality data based on the correlation. In other words, the learned neural network model identifies the main factors that determine the characteristics and properties of a drug from among a plurality of factors constituting the quality data, and based on the identified factors, the main quality profile corresponding to the quality data. can create A detailed description of the process of analyzing the correlation between the plurality of elements in the above description will be described later in detail in the learning process of the correlation analysis model.

In other words, the processor 130 may determine the elements constituting the main quality profile related to the user's input or through a pre-trained neural network model. Accordingly, a main quality profile corresponding to the corresponding quality data may be acquired through the main elements (or some items) of the quality data. For example, each of the main quality profiles related to a specific element may be obtained in correspondence with each of a plurality of quality data corresponding to various drugs.

In addition, the processor 130 may group each of the plurality of quality data into one or more data sets based on the main quality profile corresponding to each of the plurality of quality data. That is, each of the plurality of quality data may be classified into one or more data sets through grouping by the processor 130 . In other words, each of the plurality of quality data may be classified into one or more data sets based on the main quality profile. For example, the one or more data sets may include a first data set, a second data set, a third data set, and the like. In this case, each data set may be classified based on a main quality profile corresponding to each quality data. That is, one or more first quality data included in the first data set and one or more second quality data included in the second data set may be classified through different main quality profiles. As a more specific example, the product name/code of the main quality profile corresponding to one or more first quality data included in the first data set may be 'A', but one or more second quality data included in the second data set may be 'A'. The product name/code of the main quality profile corresponding to the data may be 'B'. In the foregoing description, for convenience of explanation, it is exemplarily described that grouping is performed based on 'product name/code' among a plurality of elements of the main quality profile, but various elements of the main quality profile (eg, product type/ form, indication, route of administration, process flow/unit process and manufacturing scale, etc.) will be apparent to those skilled in the art.

According to an additional embodiment, the processor 130 may encrypt each of a plurality of quality data corresponding to a plurality of medicines using a symmetric key encryption algorithm, and transmit the encrypted plurality of quality data to a separate external server (eg, cloud server). Additionally, the processor 130 may transmit a private key (or a public key) for decrypting the encrypted data to an external server. In this case, the external server may be a database server that receives and stores quality data related to various medicines, and encrypting a plurality of quality data and transmitting it to an external server may be to prevent information leakage of quality data. . That is, encrypted data cannot be read even if it is intercepted in the middle. Accordingly, confidentiality of data may be maintained during transmission and reception. When the database server receives the encrypted data and the private key (or public key), it can decrypt the data through a symmetric key decryption algorithm, and convert the decrypted data into a key quality profile corresponding to each data. can be classified based on According to an embodiment, the symmetric key encryption and decryption algorithm may be an Advanced Encryption Standard (AES-256) algorithm, which is a symmetric key algorithm that uses the same key in encryption and decryption processes. The detailed description of the above-described symmetric key algorithm is only an example, and the present disclosure is not limited thereto.

That is, data encrypted and transmitted to an external server (eg, a database server) undergoes a decryption process in the external server, and the decrypted quality data may be stored after being divided into one or more data sets based on main quality data.

In addition, the processor 130 may derive a correlation between each element included in the plurality of quality data (S130). To this end, the processor 130 performs learning on the correlation between elements constituting each of a plurality of quality data through an association rule analysis algorithm, and a correlation analysis model for deriving the correlation between the elements. can create Accordingly, the correlation between each element can be derived by using the correlation analysis model. The processor 130 may process a plurality of quality data as inputs of the correlation analysis model to derive a correlation between elements included in the quality data.

More specifically, the correlation analysis model may be a neural network model that analyzes a correlation between at least two or more elements through an association rule analysis algorithm. The association rule analysis algorithm may be an algorithm for generating a series of rules indicating whether a set between at least two or more elements frequently occurs. In an embodiment, the association rule analysis algorithm may relate to a market basket analysis algorithm that analyzes a correlation between a characteristic drug and elements included therein based on large-scale big data. The shopping cart analysis algorithm may be implemented, for example, through at least one of the Apriori algorithm, the FP-Growth algorithm, and the DHP algorithm.

In one embodiment, the processor 130 generates a plurality of quality data and a number of rules between elements included in each quality data, and based the generated rules on support, confidence, and lift. By determining this, it is possible to derive a correlation between each element. More specifically, the processor 130 may build the learning data based on each of a plurality of elements constituting the quality data. For example, it is possible to construct learning data by displaying each of a plurality of elements as columns and displaying them as true or false depending on presence or absence as a table. The processor 130 identifies the constructed learning data, generates various rules based on each of the existing elements, and utilizes a related rule analysis algorithm (or shopping cart analysis algorithm) so that the degree of support, reliability, and improvement is greater than or equal to a certain standard value. By selecting the rules, a correlation between the corresponding elements can be derived. For example, support may be defined through the probability of occurrence of a conditional clause (if, if-). Reliability is used to measure the strength of the association between elements, and can be defined through the conditional probability that a result clause will occur when a conditional clause is given. Also, the degree of improvement used to determine whether the generated rule has actual utility value may relate to the ratio of how many two events occur simultaneously compared to when each rule and the result corresponding to each rule are independent of each other. . For example, when the improvement level is 1, the conditional clause and the result may be independent of each other. This means that there is no significant correlation between rules, which may mean that there is no correlation between elements of the rule.

That is, as described above, the processor 130 may derive the correlation between elements constituting the quality data by using the correlation analysis model learned through the association rule analysis algorithm. The correlation between elements constituting the quality data may be, for example, information about precedence and precedence between elements and order. That is, the correlation may mean a precedence relationship defined through an association rule algorithm between each element for each of the quality data collected and organized for different drugs.

As a specific example, it can be derived that the composition and cell growth rate of each liquid medium required for cell culture can have a correlation with the concentration of acetate and magnesium ion that inhibit cell growth. In addition, a correlation related to whether the target protein production rate is the highest in a specific range of induction time points according to the degree of cell growth rate can be derived. The detailed description of the correlation derived between the various elements described above is only an example, and the present invention is not limited thereto.

In addition, the processor 130 may perform meta-data on a plurality of quality data to build a pharmaceutical quality database (S140). Specifically, the processor 130 may construct a drug quality database by performing metadataization on a plurality of quality data based on the correlation between each element. Metadataization may be for converting a plurality of quality data into big data. That is, the processor 130, based on the correlation between the derived elements, information on precedence, index information (eg, index information on rows and columns in DB), profile data, and verification Metadata can be formed by additionally connecting data and the like. According to this metadataization, a plurality of quality data may be converted into big data, and thus, processing efficiency may be improved in the process of searching for data similar to specific query data for a drug quality database, which will be described later. Hereinafter, a process of providing a response corresponding to the query data will be described in detail with reference to FIG. 4 .

4 is a flowchart illustrating a process of building a database by acquiring a plurality of quality data related to an embodiment of the present invention.

According to an embodiment of the present invention, the processor 130 may obtain query data (S210). The query data of the present invention may be data related to a query that occurs during drug R&D and production process. For example, the query data may be information about the quality data of a new drug under review that is not defined or stored in the existing database. As a specific example, the query data may relate to the overall conditions (eg, set temperature, humidity, experiment time, etc. related to the experimental environment) for the development of a production process and analysis method for a specific drug or quality verification experiment related to a specific drug. . For another example, the query data may include at least some of the quality characteristics of a specific drug (eg, an aggregation corresponding to a product-related impurity that is one of the quality characteristics of a drug), or a cut product ( truncated form) or a glycosylation pattern corresponding to structural properties). The detailed description of the above-described query data is only an example, and the present invention is not limited thereto.

Such query data may be obtained through a query issued from a client. In addition, the query data may be obtained as it is input through a quality document form on the eQMS. Specifically, the eQMS may include an input window for allowing the user to input a query. The user may input data related to a query occurring in the process of drug R&D and production in the corresponding input window, and the processor 130 may obtain the query data based on the corresponding input value.

According to an additional embodiment, the processor 130 may encrypt the query data using a symmetric key encryption algorithm, and transmit the encrypted query data to a separate external server (eg, a cloud server). Additionally, the processor 130 may transmit a private key (or a public key) for decrypting the encrypted data to an external server. In this case, encrypting a plurality of query data and transmitting it to an external server may be to prevent information leakage of the query data. That is, encrypted data cannot be read even if it is intercepted in the middle. Accordingly, confidentiality of data may be maintained during transmission and reception.

According to an embodiment of the present invention, the processor 130 may perform metadataization on the query data (S220). For example, the decrypted query data can be classified into a specific data set with the same main quality profile, and index information and profile data information related to rows and columns corresponding to the data set are connected to the query data as metadata. Metadataization may be performed. That is, the processor 130 may convert the decoded query data into metadata by connecting the index information on the DB, the electronic form information, and the empirical data information related to the profile data as metadata. As the query data becomes metadata, search efficiency in big data can be improved.

According to an embodiment of the present invention, the processor 130 may perform an operation on the query data (S230). The operation on the query data may be an operation for selecting quality data having a similarity greater than or equal to the query data in the drug quality database.

Specifically, the processor 130 may select one or more similar data sets 310 by performing a search on the drug quality database based on the query metadata. Here, the one or more similar data sets 310 may refer to data sets having a degree of similarity between the query data and the query data in a predetermined order or more. The processor 130 may select one or more similar data sets 310 having a similarity greater than or equal to a threshold similarity score or a major quality profile corresponding to the query data from the drug quality database. The drug quality database may be stored by grouping each of a plurality of quality data into one or more data sets based on one major quality profile (ie, CQP). That is, the processor 130 may calculate the similarity between each of the main quality profile corresponding to the query data and the main quality profile representing each data set in big data, and only the data set corresponding to the threshold similarity score or higher is one or more similar data. A set 310 may be selected. In one embodiment, the similarity comparison between the main quality profiles is vectorized for two comparison objects (that is, the main quality profile corresponding to the query data and the main quality profile corresponding to a specific data set) through functionalization on the text, and , it may be characterized in that it is performed through a cosine similarity comparison between each vector. For example, the functionalization of the text may be characterized in that it is vectorized based on the Word2Vec algorithm. For example, the calculated similarity comparison value may show a value close to 1 as the comparison target is identical or similar, and may show a value close to -1 as the comparison target is different. The detailed description of the method for comparing the degree of similarity between the texts described above is merely an example, and the present invention is not limited thereto.

That is, the processor 130 may select one or more similar data sets 310 corresponding to the query data by considering only the items corresponding to the main quality profile, rather than comparing each of the plurality of items. In other words, since the drug quality database is grouped into one or more data sets through the main quality profile, the processor 130 compares the similarity between each of the main quality profiles representing each data set and the main quality profile corresponding to the query data. Through this, quality data having a similarity greater than or equal to a certain reference value can be selected. In this case, since the number of items constituting the main quality profile is significantly smaller than that of the elements constituting the quality data, computing power consumed for calculation (ie, comparison of similarity) may be reduced. In addition, since it is a comparison with a main quality profile representing a grouped data set rather than a similarity between each of various quality data corresponding to a plurality of medicines and the query data, the amount of computation can be significantly reduced. That is, as the amount of computation is reduced in the process of selecting data similar to the query data, computational efficiency can be improved, such as by minimizing computation time.

According to an embodiment of the present invention, a threshold similarity score as a criterion for similarity comparison (or similar data set selection) may be calculated based on a key quality profile corresponding to each of one or more data sets. Specifically, the processor 130 may calculate a similarity score between key quality profiles corresponding to each of one or more data sets, and may be calculated based on one or more similarity scores generated corresponding to each data set pair. For example, when one or more similarity scores generated corresponding to each data set pair are 70, 80, 60, or 95, the processor 130 may determine the threshold similarity score to be 76.25 through an average value of the corresponding similarity scores. . Specific numerical descriptions of the above-described similarity score and threshold similarity score are only examples, and the present disclosure is not limited thereto.

The above-described threshold similarity calculation method may be for reflecting the trend of the entire data set. For example, as a result of a 1:1 similarity comparison between data sets, each similarity score may include a plurality of high scores or low scores. Accordingly, the present invention selects a threshold similarity score based on a similarity comparison between all data sets (that is, a similarity comparison between key quality profiles between each data set) (i.e., sets a threshold value for similar data selection), so that similar data The similarity trend can be reflected in the process of searching the set. Accordingly, the tendency of the entire data can be reflected in the similar data selection process for the query data, thereby ensuring improved reliability in the selection of the similar data set.

According to an embodiment of the present invention, the processor 130 may perform data grouping through classification of a plurality of quality data and query data corresponding to one or more similar data sets 310 . Specifically, the processor 130 may identify one or more elements constituting each of a plurality of quality data corresponding to one or more similar data sets. In addition, the processor 130 may perform data grouping by classifying a plurality of quality data and query data into one or more data groups, respectively, based on one or more elements.

In this case, grouping for one or more similar data sets 310 may mean regrouping 320 shown in FIG. 5 . The regrouping 320 may mean regrouping the quality data and query data grouped based on the main quality profile based on a plurality of elements constituting the respective quality data and query data. This regrouping 320 may be performed through a classification model learned based on the k-means algorithm. Specifically, the processor 130 may vectorize each element constituting the quality data, display it in an arbitrary dimensional space, and set k centroids based on an initial cluster formed by each element. After setting k centroids, the processor 130 may allocate centroids based on a distance between clusters formed by each element. In other words, each centroid may be assigned to a position close to each element. Thereafter, the processor 130 may update each centroid by moving each centroid to the center of the cluster in response to each cluster. The processor 130 may optimize the algorithm by repeating the centroid allocation and update process until the cluster allocation does not change or a predetermined tolerance or the maximum number of iterations is reached. For example, the processor 130 may perform optimization by identifying that the allowable error value for the change amount returns to within a certain level while repeatedly calculating the sum of squared errors whenever the centroid changes.

That is, data based on each value indicated by a plurality of elements of a plurality of quality data (eg, quality data included in one or more similar data sets) through the k-means algorithm-based classification model generated through the above process can be regrouped.

For example, when the one or more similar data sets 310 selected in response to the query data are the first similar data set 311 , the second similar data set 312 , and the third similar data set 313 , the processor 130 identifies all quality data included in each similar data set, and regroups 320 based on values (eg, PH, temperature at a specific stage, etc.) indicated by each element of the identified quality data. , the first data group 331 , the second data group 332 , and the third data group 333 may be formed. For example, data having similar element indication values may be classified into the same data group.

That is, the processor 130 may regroup the data classified based on the existing CQP into one or more data groups 330 based on the indicated value of each element. In this case, the query data may also be included in the regrouped data group. In other words, in the regrouping process, the query data may also be classified and grouped into the indicated data group. For example, as regrouping is performed, the query data may be classified into the first data group 331 .

According to an embodiment, the processor 130 utilizes a prediction model learned based on a K-Nearest Neighbors (KNN) algorithm to predict the data closest to or matching the query data within a data group including the query data. have. The KNN algorithm is generated through learning of a neural network based on supervised learning, and may be a neural network model that finds one of the nearest learning data points as a nearest neighbor and uses it for prediction. For example, a training data set may be formed based on a plurality of elements of quality data.

That is, the processor 130 may identify data most similar to the corresponding query data in the data group including the query data by using the knn algorithm-based prediction model.

According to an embodiment of the present invention, the processor 130 may provide drug analysis information corresponding to the query data based on the data grouping result. Specifically, the processor 130 may provide the drug analysis information by calculating the correlation of each element based on the data group into which the query data is classified.

In more detail, the processor 130 may derive a correlation between a data group in which query data is classified among one or more data groups and each of the remaining data groups by using a correlation analysis model. In this case, one or more data groups may be regrouped based on each element. In addition, the processor 130 may provide drug analysis information corresponding to the query data based on the correlation between each data group.

Here, the correlation analysis model may be a neural network model that derives the correlation between elements by learning about the correlation between elements constituting each of a plurality of quality data through an association rule analysis algorithm. .

The correlation between elements constituting the quality data may be, for example, information about precedence and precedence between elements and order. That is, the correlation may mean a precedence relationship defined through an association rule algorithm between each element for each of the quality data collected and organized for different drugs.

As a specific example, it can be derived that the composition and cell growth rate of each liquid medium required for cell culture can have a correlation with the concentration of acetate and magnesium ion that inhibit cell growth. In addition, a correlation related to whether the target protein production rate is the highest in a specific range of induction time points according to the degree of cell growth rate can be derived. The detailed description of the correlation derived between the various elements described above is only an example, and the present invention is not limited thereto.

That is, the processor 130 can analyze the correlation between the data group into which the query data is classified and other data groups by using the correlation analysis model learned in advance to derive the correlation between each element included in the plurality of quality data. And based on the analyzed correlation, drug analysis information can be provided. In other words, the correlation between the data group containing the query data and other data groups can be derived through the correlation analysis model. For example, when the first data group to which the query data belongs is related to 'aggregation', various data groups may be sorted and provided according to the order in which the degree of correlation with the first group is high. In one example, when the user selects the second data group (in-process control or shipment test) from among the sorted data groups, drug analysis information may be generated and provided based on values included in the second data group. .

For a more specific example, when query data of 'aggregation' is obtained, the processor 130 uses the main quality profile for the candidate material corresponding to the query data as an index, quality risk analysis data, quality characteristics, and product-related impurities Profile data such as profile data and empirical data including permissible ranges can be meta-dataized. In addition, the processor 130 selects one or more data sets having a major quality profile that meets the degree of similarity above the standard from the drug quality database, and regroups the data sets to the data group corresponding to the 'quality risk analysis data' element. Drug analysis information can be provided by classifying the query data and determining that 'aggregation' corresponds to a high risk through correlation analysis between the data group into which the query data is classified and other data groups. The detailed description of the above-described drug analysis information providing process is only an example, and the present disclosure is not limited thereto.

The drug analysis information of the present invention is information about a response to a user's inquiry related to the development and production of a drug, and may include at least one of drug-related element information, drug tendency information, and drug process risk information.

Here, the drug relation element information relates to the input data that causes the predicted result data and the optimal result data in response to the query data when designing the experimental method, and the adjustment data that suggests adjustment of the relationship between each element. may contain information. For example, when developing a monoclonal antibody, drug-related element information includes information on antibody-dependent cellular cytotoxicity (ADCC) and fucosylation level as major quality characteristics for candidate substances. may include As an example, average partial CO2 level may be suggested as critical process parameters (CPPs) to provide a process control range of the corresponding variable. The detailed description of the above-described drug-related element information is only an example, and the present disclosure is not limited thereto.

In addition, the drug trend information may include information on warning indications and follow-up measures for the possibility of a deviation from a standard or a deviation from a trend when quality trend analysis is performed on the trend element. For example, when producing insulin analogs, drug trend information indicates that the concentration of residual c-peptide in the DS release QC testing of the candidate substance tends to be constant. It is possible to predict the possibility of occurrence of deviation beyond the upper limit while showing, and include information on the timing of occurrence of deviation and cause of occurrence of deviation. For example, the column recycle cycle of ion exchange-HPLC, the second purification chromatography that may be involved in residual c-peptide, can be analyzed as a cause, and new column packing can be proposed as a solution. have. The detailed description of the drug tendency information described above is only an example, and the present disclosure is not limited thereto.

In addition, the drug process risk information may be information on a high-risk hazard factor among major factors for each process step in the production process step and information on a mitigation plan for reducing the high risk. For example, when using certain equipment in the finished process of biopharmaceutical production, drug process risk information informs about frequent deviations in the vial cleaning process, and in order to prevent this, This may include information suggesting a mitigation plan, such as a shorter replacement cycle.

As described above, the processor 130 builds a drug quality database based on a large amount of electronic quality data for the entire drug through eQMS, and conducts R&D and It can provide responses (or answers) to various issues that occur in the production process. Since this is based on big data data, that is, correlation analysis between components of various drugs, meaningful prediction information can be provided to users in the development, research, and production stages.

6 is a flowchart exemplarily illustrating a method of providing a response to inquiry data based on quality data of medicines related to an embodiment of the present invention.

According to an embodiment of the present invention, the method may include obtaining query data (S310).

According to an embodiment of the present invention, the method may include performing metadataization on the query data ( S320 ).

According to an embodiment of the present invention, the method may include selecting one or more similar data sets by performing a search on a drug quality database based on the query data (S330).

According to an embodiment of the present invention, the method may include performing data grouping through classification of a plurality of quality data and query data corresponding to one or more similar data sets ( S340 ).

According to an embodiment of the present invention, the method may include providing (S350) drug analysis information corresponding to the query data based on the data grouping result.

The order of the steps illustrated in FIG. 6 described above may be changed if necessary, and at least one or more steps may be omitted or added. That is, the above-described steps are merely an embodiment of the present invention, and the scope of the present invention is not limited thereto.

7 is a schematic diagram illustrating one or more network functions related to an embodiment of the present invention.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as “nodes”. These “nodes” may also be referred to as “neurons”. A neural network is configured by including at least one or more nodes. Nodes (or neurons) constituting neural networks may be interconnected by one or more “links”.

In the neural network, one or more nodes connected through a link may relatively form a relationship between an input node and an output node. The concepts of an input node and an output node are relative, and any node in an output node relationship with respect to one node may be in an input node relationship in a relationship with another node, and vice versa. As described above, an input node-to-output node relationship may be created around a link. One or more output nodes may be connected to one input node through a link, and vice versa.

In the relationship between the input node and the output node connected through one link, the value of the output node may be determined based on data input to the input node. Here, a node interconnecting the input node and the output node may have a weight. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the weight.

As described above, in a neural network, one or more nodes are interconnected through one or more links to form an input node and an output node relationship in the neural network. The characteristics of the neural network may be determined according to the number of nodes and links in the neural network, the correlation between the nodes and the links, and the value of a weight assigned to each of the links. For example, when the same number of nodes and links exist and there are two neural networks having different weight values between the links, the two neural networks may be recognized as different from each other.

A neural network may include one or more nodes. Some of the nodes constituting the neural network may configure one layer based on distances from the initial input node. For example, a set of nodes having a distance of n from the initial input node is You can configure n layers. The distance from the initial input node may be defined by the minimum number of links that must be passed to reach the corresponding node from the initial input node. However, the definition of such a layer is arbitrary for description, and the order of the layer in the neural network may be defined in a different way from the above. For example, a layer of nodes may be defined by a distance from the final output node.

The initial input node may mean one or more nodes to which data is directly input without going through a link in a relationship with other nodes among nodes in the neural network. Alternatively, in a relationship between nodes based on a link in a neural network, it may mean nodes that do not have other input nodes connected by a link. Similarly, the final output node may refer to one or more nodes that do not have an output node in relation to other nodes among nodes in the neural network. In addition, the hidden node may mean nodes constituting the neural network other than the first input node and the last output node. The neural network according to an embodiment of the present invention may be a neural network in which the number of nodes in the input layer may be the same as the number of nodes in the output layer, and the number of nodes decreases and then increases again as the input layer progresses to the hidden layer. can In addition, in the neural network according to another embodiment of the present invention, the number of nodes in the input layer may be less than the number of nodes in the output layer, and the number of nodes may be reduced as the number of nodes progresses from the input layer to the hidden layer. have. In addition, the neural network according to another embodiment of the present invention may be a neural network in which the number of nodes in the input layer may be greater than the number of nodes in the output layer, and the number of nodes increases as the number of nodes progresses from the input layer to the hidden layer. can The neural network according to another embodiment of the present invention may be a neural network in the form of a combination of the aforementioned neural networks.

A deep neural network (DNN) may refer to a neural network including a plurality of hidden layers in addition to an input layer and an output layer. Deep neural networks can be used to identify the latent structures of data. In other words, it can identify the potential structure of photos, texts, videos, voices, and music (e.g., what objects are in the photos, what the text and emotions are, what the texts and emotions are, etc.) . Deep neural networks include convolutional neural networks (CNNs), recurrent neural networks (RNNs), auto encoders, generative adversarial networks (GANs), and restricted boltzmann machines (RBMs). machine), a deep trust network (DBN), a Q network, a U network, a Siamese network, and the like. The description of the deep neural network described above is only an example, and the present invention is not limited thereto.

The neural network may be learned by at least one of teacher learning (supervised learning), unsupervised learning (unsupervised learning), and semi-supervised learning (semi-supervised learning). The training of the neural network is to minimize the error in the output. In the training of a neural network, iteratively input the training data into the neural network, calculate the output of the neural network and the target error for the training data, and calculate the error of the neural network from the output layer of the neural network to the input layer in the direction to reduce the error. It is a process of updating the weight of each node in the neural network by backpropagation in the direction. In the case of teacher learning, learning data in which the correct answer is labeled in each learning data is used (ie, labeled learning data), and in the case of comparative learning, the correct answer may not be labeled in each learning data. That is, for example, learning data in the case of teacher learning related to data classification may be data in which categories are labeled in each of the learning data. Labeled training data is input to the neural network, and an error can be calculated by comparing the output (category) of the neural network with the label of the training data. As another example, in the case of comparison learning related to data classification, an error may be calculated by comparing the input training data with the neural network output. The calculated error is backpropagated in the reverse direction (ie, from the output layer to the input layer) in the neural network, and the connection weight of each node of each layer of the neural network may be updated according to the backpropagation. A change amount of the connection weight of each node to be updated may be determined according to a learning rate. The computation of the neural network on the input data and the backpropagation of errors can constitute a learning cycle (epoch). The learning rate may be applied differently according to the number of repetitions of the learning cycle of the neural network. For example, in the early stage of learning of a neural network, a high learning rate can be used to enable the neural network to quickly obtain a certain level of performance, thereby increasing efficiency, and using a low learning rate at a later stage of learning can increase accuracy.

In the training of neural networks, in general, the training data may be a subset of real data (that is, data to be processed using the trained neural network), and thus, the error on the training data is reduced, but the error on the real data is reduced. There may be increasing learning cycles. Overfitting is a phenomenon in which errors on actual data increase by over-learning on training data as described above. For example, a phenomenon in which a neural network that has learned a cat by seeing a yellow cat does not recognize that it is a cat when it sees a cat other than yellow may be a type of overfitting. Overfitting can act as a cause of increasing errors in machine learning algorithms. In order to prevent such overfitting, various optimization methods can be used. In order to prevent overfitting, methods such as increasing training data, regularization, or dropout in which a part of nodes in the network are omitted in the process of learning, may be applied.

Throughout this specification, computational model, neural network, network function, and neural network may be used interchangeably. (Hereinafter, the neural network is unified and described.) The data structure may include a neural network. And the data structure including the neural network may be stored in a computer-readable medium. Data structures, including neural networks, may also include data input to the neural network, weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, activation functions associated with each node or layer of the neural network, and loss functions for learning the neural network. have. A data structure comprising a neural network may include any of the components disclosed above. That is, the data structure including the neural network includes all or all of the data input to the neural network, the weights of the neural network, hyperparameters of the neural network, data obtained from the neural network, the activation function associated with each node or layer of the neural network, and the loss function for training the neural network. may be configured including any combination of In addition to the above-described configurations, a data structure including a neural network may include any other information that determines a characteristic of a neural network. In addition, the data structure may include all types of data used or generated in the operation process of the neural network, and is not limited to the above. Computer-readable media may include computer-readable recording media and/or computer-readable transmission media. A neural network may be composed of a set of interconnected computational units, which may generally be referred to as nodes. These nodes may also be referred to as neurons. A neural network is configured by including at least one or more nodes.

The data structure may include data input to the neural network. A data structure including data input to the neural network may be stored in a computer-readable medium. The data input to the neural network may include learning data input in a neural network learning process and/or input data input to the neural network in which learning is completed. Data input to the neural network may include pre-processing data and/or pre-processing target data. The preprocessing may include a data processing process for inputting data into the neural network. Accordingly, the data structure may include data to be pre-processed and data generated by pre-processing. The above-described data structure is merely an example, and the present invention is not limited thereto.

The data structure may include the weights of the neural network. (In this specification, weight and parameter may be used interchangeably.) And the data structure including the weight of the neural network may be stored in a computer-readable medium. The neural network may include a plurality of weights. The weight may be variable, and may be changed by a user or an algorithm in order for the neural network to perform a desired function. For example, when one or more input nodes are interconnected to one output node by respective links, the output node sets values input to input nodes connected to the output node and links corresponding to the respective input nodes. An output node value may be determined based on the parameter. The above-described data structure is merely an example, and the present invention is not limited thereto.

By way of example and not limitation, the weight may include a weight variable in a neural network learning process and/or a weight in which neural network learning is completed. The variable weight in the neural network learning process may include a weight at a time point at which a learning cycle starts and/or a weight variable during the learning cycle. The weight for which neural network learning is completed may include a weight for which a learning cycle has been completed. Accordingly, the data structure including the weights of the neural network may include a data structure including the weights that vary in the process of learning the neural network and/or the weights on which the learning of the neural network is completed. Therefore, it is assumed that the above-described weights and/or combinations of weights are included in the data structure including the weights of the neural network. The above-described data structure is merely an example, and the present invention is not limited thereto.

The data structure including the weights of the neural network may be stored in a computer-readable storage medium (eg, memory, hard disk) after being serialized. Serialization can be the process of converting a data structure into a form that can be reconstructed and used later by storing it on the same or a different computing device. The computing device may serialize the data structure to send and receive data over the network. A data structure including weights of the serialized neural network may be reconstructed in the same computing device or in another computing device through deserialization. The data structure including the weight of the neural network is not limited to serialization. Furthermore, the data structure including the weights of the neural network is a data structure to increase computational efficiency while using the resources of the computing device to a minimum (e.g., B-Tree, Trie, m-way search tree, AVL tree, Red-Black Tree). The foregoing is merely an example, and the present invention is not limited thereto.

The data structure may include hyper-parameters of the neural network. In addition, the data structure including the hyperparameters of the neural network may be stored in a computer-readable medium. The hyper parameter may be a variable variable by a user. Hyperparameters are, for example, learning rate, cost function, number of iterations of the learning cycle, weight initialization (e.g., setting the range of weight values to be initialized for weights), Hidden Unit The number (eg, the number of hidden layers, the number of nodes of the hidden layer) may be included. The above-described data structure is merely an example, and the present invention is not limited thereto.

The steps of a method or algorithm described in relation to an embodiment of the present invention may be implemented directly in hardware, as a software module executed by hardware, or by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any type of computer-readable recording medium well known in the art to which the present invention pertains.

The components of the present invention may be implemented as a program (or application) to be executed in combination with a computer, which is hardware, and stored in a medium. Components of the present invention may be implemented as software programming or software components, and similarly, embodiments may include various algorithms implemented as data structures, processes, routines, or combinations of other programming constructs, including C, C++ , Java, assembler, etc. may be implemented in a programming or scripting language. Functional aspects may be implemented in an algorithm running on one or more processors.

A person of ordinary skill in the art will recognize that the various illustrative logical blocks, modules, processors, means, circuits and algorithm steps described in connection with the embodiments disclosed herein are implemented in electronic hardware, (convenience For this purpose, it will be understood that it may be implemented by various forms of program or design code (referred to herein as "software") or a combination of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. A person skilled in the art may implement the described functionality in various ways for each specific application, but such implementation decisions should not be interpreted as a departure from the scope of the present invention.

The various embodiments presented herein may be implemented as methods, apparatus, or articles of manufacture using standard programming and/or engineering techniques. The term “article of manufacture” includes a computer program, carrier, or media accessible from any computer-readable device. For example, computer-readable media include magnetic storage devices (eg, hard disks, floppy disks, magnetic strips, etc.), optical disks (eg, CDs, DVDs, etc.), smart cards, and flash memory. devices (eg, EEPROMs, cards, sticks, key drives, etc.). Also, various storage media presented herein include one or more devices and/or other machine-readable media for storing information. The term “machine-readable medium” includes, but is not limited to, wireless channels and various other media that can store, hold, and/or convey instruction(s) and/or data.

It is to be understood that the specific order or hierarchy of steps in the presented processes is an example of exemplary approaches. Based on design priorities, it is to be understood that the specific order or hierarchy of steps in the processes may be rearranged within the scope of the present invention. The appended method claims present elements of the various steps in a sample order, but are not meant to be limited to the specific order or hierarchy presented.

The description of the presented embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the scope of the invention. Thus, the present invention is not to be limited to the embodiments presented herein but should be construed in the widest scope consistent with the principles and novel features presented herein.

Claims (10)

A method performed on one or more processors of a computing device, comprising:
establishing a drug quality database;
obtaining query data;
selecting one or more similar data sets by performing a search on the drug quality database based on the query data;
performing data grouping through classification of a plurality of quality data corresponding to the one or more similar data sets and the query data; and
providing drug analysis information corresponding to the query data based on the data grouping result;
includes,
The step of constructing the drug quality database is,
acquiring a plurality of quality data (OQD, Overall Quality Data) corresponding to each of a plurality of pharmaceuticals; and
grouping each of the plurality of quality data into one or more data sets based on a Critical Quality Profile (CQP) corresponding to each of the plurality of quality data;
containing,
A method of providing responses to inquiry data based on drug quality data. delete According to claim 1,
The quality data is
It includes general information of drugs, profile data related to production and quality definition, and empirical data corresponding to the profile data,
The main quality profile is,
Information about the main factors that can determine the characteristics and properties of the drug, which is utilized for the search, and is configured through at least a part of the quality data,
A method of providing responses to inquiry data based on drug quality data. According to claim 1,
The step of constructing the drug quality database is,
generating a correlation analysis model for deriving a correlation between each element by learning about a correlation between elements constituting each of the plurality of quality data through a correlation rule analysis algorithm; and
constructing the pharmaceutical quality database by performing meta-dataization on the plurality of quality data based on the correlation between the respective elements;
containing,
A method of providing responses to inquiry data based on drug quality data. A method performed on one or more processors of a computing device, comprising:
obtaining query data;
selecting one or more similar data sets by performing a search on a drug quality database based on the query data;
performing data grouping through classification of a plurality of quality data corresponding to the one or more similar data sets and the query data; and
providing drug analysis information corresponding to the query data based on the data grouping result;
includes,
The step of selecting one or more similar data sets comprises:
selecting one or more similar data sets having a similarity greater than or equal to a critical similarity score to a major quality profile corresponding to the query data from the drug quality database;
including,
The critical similarity score is
Computing the similarity score between the main quality profiles corresponding to each of the one or more data sets, characterized in that it is calculated based on the one or more similarity scores generated corresponding to the respective data set pairs,
A method of providing responses to inquiry data based on drug quality data. delete 5. The method of claim 4,
performing data grouping through classification of a plurality of quality data corresponding to the one or more similar data sets and the query data,
identifying one or more elements constituting each of a plurality of quality data corresponding to the one or more similar data sets; and
performing data grouping by classifying each of the plurality of quality data and the query data into one or more data groups based on the one or more elements;
containing,
A method of providing responses to inquiry data based on drug quality data. 8. The method of claim 7,
The step of providing the drug analysis information,
deriving a correlation between a data group in which the query data is classified among the one or more data groups and each of the remaining data groups by using the correlation analysis model; and
providing drug analysis information corresponding to the query data based on the correlation between the respective data groups;
containing,
A method of providing responses to inquiry data based on drug quality data. a storage unit for storing one or more instructions; and
a processor executing one or more instructions stored in the storage unit; including,
The processor by executing the one or more instructions,
A computing device for performing the method of claim 1 or 5 . A computer program stored in a computer-readable recording medium in combination with a computer, which is hardware, to perform the method of claim 1 or 5.

Images