CN119476883A

CN119476883A - Intelligent decision-making method, device, equipment and medium based on business configuration

Info

Publication number: CN119476883A
Application number: CN202510052513.5A
Authority: CN
Inventors: 丁新云; 杨作銘; 叶惠陆
Original assignee: Eden Information Service Ltd
Current assignee: Eden Information Service Ltd
Priority date: 2025-01-14
Filing date: 2025-01-14
Publication date: 2025-02-18

Abstract

The application discloses an intelligent decision method, a device, equipment and a medium based on service configuration, wherein the intelligent decision method based on service configuration comprises the steps of acquiring real-time service data in a service system, adopting a machine learning algorithm to learn and identify noise data characteristics, dynamically updating a data cleaning rule, cleaning the real-time service data according to the updated data cleaning rule to obtain standardized service data, acquiring a current service configuration model according to a service type, importing the standardized service data into the current service configuration model, analyzing and processing the standardized service data through the current service configuration model, and generating an intelligent decision result corresponding to the service type through a decision logic module.

Description

Intelligent decision method, device, equipment and medium based on service configuration

Technical Field

The present invention relates to the field of data processing technologies, and in particular, to an intelligent decision method, apparatus, device, and medium based on service configuration.

Background

Currently, the scene service required by various industries is increasingly complicated and diversified in the face of massive data and transient business environments. The traditional service decision mode mostly depends on manual experience, and a decision maker needs to spend a great deal of time to collect, sort and analyze various service data, and try to find rules from complicated information so as to make proper decisions. For example, in the field of supply chain management, many factors such as raw material price fluctuation, supplier delivery cycle, stock level, market demand prediction and the like are considered manually, so that the efficiency is low, and decision errors are easily caused by human negligence or cognitive limitation. Even if part of business scenes are introduced into information automation system auxiliary decisions, the systems are often poor in flexibility and run against preset scenes based on fixed algorithms. When the business rule, flow or external environment changes, the original automatic decision module can not quickly and accurately adjust the decision according to the new situation, so that the new requirement caused by business change is difficult to be met adaptively.

Disclosure of Invention

The embodiment of the invention provides an intelligent decision method, device, equipment and medium based on service configuration, which are used for solving the problem that an original automatic decision module cannot quickly and accurately adjust decision according to new conditions, so that new requirements caused by service variation are difficult to adaptively meet.

An intelligent decision method based on service configuration, comprising:

Collecting real-time service data in a service system, wherein the real-time service data comprises service types, service flow information, service resource state information and user interaction information;

The noise data characteristics are learned and identified by adopting a machine learning algorithm, a data cleaning rule is dynamically updated, real-time service data is cleaned according to the updated data cleaning rule, and the noise data and redundant data are removed, so that standardized service data are obtained;

According to the service type, a matched historical service configuration model is determined, real-time fine adjustment is carried out on the service configuration model based on real-time data, and a current service configuration model is obtained;

importing the standardized service data into a current service configuration model;

And analyzing and processing the standardized service data through the current service configuration model, and generating an intelligent decision result corresponding to the service type through the decision logic module.

An intelligent decision making device based on service configuration, comprising:

The service data acquisition module is used for acquiring real-time service data in the service system, wherein the real-time service data comprises service types, service flow information, service resource state information and user interaction information;

The service data obtaining module is used for learning and identifying noise data characteristics by adopting a machine learning algorithm, dynamically updating a data cleaning rule, cleaning real-time service data according to the updated data cleaning rule, and removing the noise data and redundant data to obtain standardized service data;

The service model acquisition module is used for determining a matched historical service configuration model according to the service type, and carrying out real-time fine adjustment on the service configuration model based on real-time data to acquire a current service configuration model;

The service data importing module is used for importing standardized service data into the current service configuration model;

And the decision result generation module is used for analyzing and processing the standardized service data through the current service configuration model and generating an intelligent decision result corresponding to the service type through the decision logic module.

An electronic device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the above-described business configuration based intelligent decision method when executing the computer program.

A computer readable medium storing a computer program which when executed by a processor implements the intelligent decision method based on service configuration described above.

According to the intelligent decision method, the device, the equipment and the medium based on service configuration, the real-time service data in the service system are collected, the noise data characteristics are learned and identified by adopting the machine learning algorithm, the data cleaning rule is dynamically updated, the real-time service data are cleaned according to the updated data cleaning rule to obtain the standardized service data, the current service configuration model is obtained according to the service type, the standardized service data are imported into the current service configuration model, the standardized service data are analyzed and processed through the current service configuration model, and the intelligent decision result corresponding to the service type is generated through the decision logic module.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram illustrating an application environment of an intelligent decision method based on service configuration according to an embodiment of the present invention;

FIG. 2 is a first flowchart of an intelligent decision method based on service configuration according to a first embodiment of the present invention;

FIG. 3 is a second flowchart of an intelligent decision method based on service configuration according to a second embodiment of the present invention;

FIG. 4 is a third flowchart of an intelligent decision method based on service configuration according to a third embodiment of the present invention;

FIG. 5 is a fourth flowchart of an intelligent decision method based on service configuration according to a fourth embodiment of the present invention;

FIG. 6 is a schematic diagram of an intelligent decision device based on service configuration according to an embodiment of the invention;

Fig. 7 is a schematic diagram of an electronic device according to an embodiment of the invention.

Detailed Description

The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

The intelligent decision method based on the service configuration provided by the embodiment of the invention can be applied to an application environment as shown in fig. 1, and is applied to an intelligent decision system based on the service configuration, wherein the intelligent decision system based on the service configuration comprises a client and a server, and the client communicates with the server through a network. The client is also called a client, and refers to a program corresponding to a server and providing local services for the client.

Further, the client is a computer-side program, an APP program of the intelligent device or a third party applet embedded with other APP. The client may be installed on, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers, portable wearable devices, and other electronic devices.

The server can be realized by an independent server or a server cluster formed by a plurality of servers, an infrastructure of the whole intelligent decision system is built, and an initial service configuration model framework comprising a service rule base, a service parameter base and a decision logic module is built. The business rule base sorts and stores detailed operation rules of various businesses according to past business experience, industry specifications and enterprise internal standards, the business parameter base refers to key parameters in business processes, such as an order processing time upper limit, a resource utilization rate threshold value and the like, and defines a conventional threshold range, and the decision logic module preliminarily has the capability of generating decision suggestions based on preset business rules and parameter threshold ranges, so that a basic framework is provided for follow-up accurate decisions based on real-time data.

In one embodiment, as shown in fig. 2, an intelligent decision method based on service configuration is provided, and the method is applied to the server in fig. 1 for illustration, and specifically includes the following steps:

S110, collecting real-time service data in a service system, wherein the real-time service data comprises service types, service flow information, service resource state information and user interaction information.

Specifically, the embodiment provides a distributed data acquisition architecture with innovation and high efficiency for ensuring that real-time information of dynamic changes in the operation process of a service system can be accurately and comprehensively captured. The architecture integrates the Internet of things sensing technology, the intelligent network routing strategy and the self-adaptive data transmission protocol. Whether it is the source end point of service start, such as the entrance interface of customer submitting service application, or the core hub of service circulation, such as the department system responsible for key approval and data processing in the internal flow, or the boundary port interacting with the external association party, such as the data interface of the interfacing supplier and partner, belongs to the information collection range.

And (3) carrying out real-time monitoring on service resource state information, taking a large cloud computing data center as an example, and continuously monitoring real-time conditions of various resources by a probe. The hardware level is used for closely focusing on the space occupancy rate of a physical cabinet of a server machine room and the real-time operation parameters of a server heat dissipation system at any time, ensuring the stable and proper operation environment of the server, monitoring the port utilization rate, the bandwidth throughput and other key indexes of network equipment such as a network switch, a router and the like in real time, ensuring the high-speed smooth data transmission, and continuously monitoring the number of various running virtual instances in a cloud Computing Management System (CMS), the length of a mass data task queue being processed, the data interaction delay condition in butt joint with an external client, the allocation utilization rate of storage resources and the like.

The acquisition of the user interaction information is illustrated through a game platform scene, and each time of game login, role creation, stage selection, prop purchase and use and communication speaking in a game community of a user, the duration and the position of the long-time stay hesitation in a game main interface are accurately captured by a probe. By means of a deep learning algorithm, an emotion analysis engine and a user behavior clustering model, game preferences of a user, such as the type of an interesting game, whether the game is competitive, role playing or leisure and intelligence improving, and capturing game emotions of the user in real time, such as frustrated emotion after continuous game transmission or excited state after high-difficulty level attack.

In summary, through the omnibearing, real-time and deep acquisition of service types, service flow information, service resource state information and user interaction information, effective data is provided for the intelligent decision-making system of the embodiment of the application, so that the following key flows such as data cleaning, dynamic model optimization, accurate decision-making and the like are driven to operate efficiently.

S120, learning and identifying noise data features by adopting a machine learning algorithm, dynamically updating a data cleaning rule, cleaning real-time service data according to the updated data cleaning rule, and removing the noise data and redundant data to obtain standardized service data.

Specifically, the machine learning algorithm is not limited to algorithms such as a support vector machine, a random forest, a gradient lifting tree and the like, and is used for carrying out deep learning identification on noise data characteristics so as to adapt to dynamic changes of service data. With the continuous promotion of business and the continuous accumulation of data, the algorithm continuously evolves and the data cleaning rule is dynamically updated. According to the updated rule, the service data acquired in real time is subjected to fine cleaning, noise data is accurately removed, related data is extracted from a data source, cleaning, conversion and standardization processing are performed, so that decision input data such as error data generated by network fluctuation, abnormal readings caused by sensor faults and redundant data such as repeatedly recorded service flow information are formed, and therefore high-quality standardized service data is obtained, and effective basic data is provided for subsequent analysis processing.

S130, determining a matched historical service configuration model according to the service type, and performing real-time fine adjustment on the service configuration model based on real-time data to obtain a current service configuration model.

Specifically, according to the service type in the collected real-time service data, a history service configuration model matched with the service type is rapidly determined in an existing model library. Meanwhile, based on the real-time property, complexity and data volume characteristics of the real-time data, the model is subjected to real-time fine adjustment. The method comprises the steps of analyzing the complexity of real-time business data, properly increasing the complexity of a model if complex businesses of multi-process intersection and multi-department cooperation are involved, optimizing a model structure to improve operation efficiency for businesses with extremely high response speed by considering real-time requirements, and reasonably expanding the processing capacity of the model under a large-data-volume scene by combining the size of data volume. And introducing a model library containing a pre-trained model template, and rapidly generating a current service configuration model which is accurately adapted to the current service requirement according to the model type and complexity determined by the analysis.

S140, importing the standardized service data into a current service configuration model.

S150, the current service configuration model analyzes and processes the standardized service data, and an intelligent decision result corresponding to the service type is generated through the decision logic module.

Specifically, the standardized service data includes service rules, user-defined policies, and system default policies, and the service rules, user-defined policies, and system default policies are combined to generate a comprehensive service rule, which is parsed to generate a logic expression, and the logic expression is converted into an executable decision tree or rule engine.

The current service configuration model predicts, classifies or optimizes the standardized service data, generates intelligent decision results corresponding to the service types through the decision logic module, and outputs final decision results in a form understandable to users, including but not limited to reports, charts or direct operation instructions. And accurately importing the cleaned standardized service data into a current service configuration model, wherein all components in the model work cooperatively. The business rule base provides basic rule guidance, the business parameter base sets key parameter constraint, the decision logic module carries out deep analysis processing on standardized business data according to core elements, an intelligent decision result aiming at the current business type is generated, and an adaptive action scheme is provided for business operation.

According to the intelligent decision method based on service configuration, through collecting real-time service data in a service system, learning and identifying noise data features by adopting a machine learning algorithm, dynamically updating a data cleaning rule, cleaning the real-time service data according to the updated data cleaning rule to obtain standardized service data, acquiring a current service configuration model according to a service type, importing the standardized service data into the current service configuration model, analyzing and processing the standardized service data through the current service configuration model, and generating an intelligent decision result corresponding to the service type through a decision logic module.

In a specific embodiment, in step S110, that is, collecting real-time service data in a service system, the method specifically includes the following steps:

s210, setting data acquisition probes at a plurality of key nodes of a service system by adopting a distributed data acquisition architecture, and summarizing acquired data to a data buffer pool in an asynchronous transmission mode.

Specifically, a distributed data acquisition architecture is adopted, data acquisition probes are deployed at each key node of a service system, and the nodes cover the starting end, key processing links, parts frequently interacted with users and the like of a service flow, so that various information of service operation can be comprehensively captured. And the collected data is efficiently and stably summarized to a data buffer pool in an asynchronous transmission mode, so that sufficient original materials are provided for subsequent data processing.

In the daily operation process of the service system, the data acquisition probes are continuously active, and real-time service data are accurately acquired. The real-time business data not only covers the business types and clearly identifies the specific categories of the current business, such as e-commerce order processing, financial loan approval and the like, but also comprises detailed business flow information, business circulation states, processing time and the like of all links, business resource state information is also of great importance, server loads, inventory numbers, human resource allocation conditions and the like are reflected in real time, and user interaction information such as operation records, feedback comments, preference selections and the like of users is also included. Through the omnidirectional acquisition, rich and three-dimensional data support is provided for intelligent decision-making, so that timeliness and comprehensiveness of data acquisition are ensured, and interference to normal operation of a service system is reduced.

In a specific embodiment, as shown in fig. 3, the intelligent decision method based on service configuration provided in this embodiment further specifically includes the following steps:

S310, carrying out periodic updating maintenance on the service configuration model, wherein the method comprises the following steps:

s311, collecting new historical service data every preset time period, updating and perfecting the service rules in the service rule base, and supplementing the new service rules or correcting the old rules.

S312, according to business development requirements and industry dynamic adjustment, recalibrating key parameters and threshold ranges thereof in a business parameter library.

S313, optimizing and training the decision logic module by using a reinforcement learning algorithm.

Specifically, a comprehensive model update maintenance flow is started every preset time period. The method comprises the steps of comprehensively collecting new historical business data, mining new business rules in the new historical business data, timely supplementing the new business rules into a business rule base, correcting old rules, for example, correcting data which does not accord with actual business conditions to ensure the real-time performance of the rules, and simultaneously, tightly combining business development requirements and industry dynamic adjustment, and recalibrating key parameters and threshold ranges thereof in a business parameter base to enable the key parameters and the threshold ranges to be matched with the current business environment. The decision logic module is subjected to high-strength optimization training by using the reinforcement learning algorithm, so that the scientificity and accuracy of decision are continuously improved, and more accurate and efficient decision suggestions can be generated based on updated business rules and parameter threshold ranges.

In a specific embodiment, before step S130, that is, before determining the matched historical service configuration model, the method further specifically includes the following steps:

S410, an initial business configuration model is built based on historical business data and an industry standard business process and used for generating the historical business configuration model based on real-time data and the initial business configuration model, wherein the initial business configuration model comprises a business rule base, a business parameter base and a decision logic module, wherein the business rule base stores operation rules of various businesses, the business parameter base stores key parameters and threshold ranges thereof in the business process, and the decision logic module generates decision suggestions according to the business rules and the parameter threshold ranges.

Specifically, the business rule base includes business rules, user-defined policies, and system default policies. In the embodiment, massive and representative historical service data are collected first, and the data are derived from service records accumulated in a long time in the past and cover practical data in different stages and under different service scenes. Taking a financial credit service as an example, historical service data comprises a loan application record of the past year, and relates to detailed information such as credit qualification, loan amount, repayment deadline, default conditions and the like of different customer groups, wherein an industry standard service flow comprises standard operation sequence and standard requirements of various links such as submitting a customer loan application, evaluating credit, checking risk, paying and approving, managing after loan and the like.

Based on the consolidated data and the flow basis, an initial service configuration model is built. Aiming at financial credit business, a rule base stores minimum credit score requirements corresponding to different loan line intervals, a loan issuing rule, a post-loan monitoring rule, a time node and a frequency of overdue repayment reminding, a triggering condition for taking a collection measure and other various operation rules, wherein the minimum credit score requirements are definitely specified by a customer qualification auditing rule, and the loan issuing rule records under which risk evaluation level can approve the loan and the upper limit of the loan line.

The business parameter store stores key parameters in the business process and threshold ranges thereof. Taking credit service as an example, key parameters include annual rate of loan, reasonable threshold ranges are set according to factors such as market interest fluctuation trend, enterprise fund cost, risk premium and the like, for example, in a market stationary period, the annual rate threshold of conventional consumption loan is set to 4% -12%, the upper limit of the annual rate of loan is set, and in combination with enterprise fund storage and risk bearing capacity, different upper limits of the annual rate are set for customers with different credit grades, such as 50 ten thousand yuan highest credit of credit high-quality customers and 20 ten thousand yuan common customers, and the weight of risk assessment indexes, such as credit score 40%, property liability 30%, income stability 30% and the like, so that the influence degree of each factor on risk judgment is accurately quantized.

The decision logic module is a decision suggestion module generated according to the business rules and the parameter threshold range. When real-time business data flows in, for example, a customer submits a loan application, the module rapidly invokes the business rule base to check whether the customer meets the basic application condition, and then evaluates the risk condition by referring to the business parameter base. If the credit score of the client is 70, applying for 30 ten thousand yuan 3 years loan, and the module judges that the risk of the client is controllable according to the set rule and parameters and combining the current market interest rate of 6%, and recommends approval of paying, and simultaneously gives out detailed decision contents such as the recommended paying interest rate of 6%, the repayment period of 3 years and the like.

The constructed initial service configuration model can be quickly fused and interacted with data after the real-time data is received, and a history service configuration model highly adapted to the current service scene is generated through continuous iterative optimization, so that intelligent decision is continuously provided.

For other industry fields, such as manufacturing industry, electronic commerce and the like, according to the characteristics of each industry, targeted optimization can be performed in the aspects of historical service data selection, industry standard flow compliance, service rule customization, parameter threshold setting, decision logic design and the like, so that the universality and the professional unification of an initial service configuration model are ensured.

In a specific embodiment, the real-time traffic data further includes a target optimization index. As shown in fig. 4, after step S150, that is, after the intelligent decision result corresponding to the service type is generated by the decision logic module, the method specifically further includes the following steps:

s510, evaluating the intelligent decision result according to the target optimization index, and extracting a parameter deviation item if the intelligent decision result does not meet the preset standard.

Specifically, the present embodiment may set a target optimization index system for different traffic types. Taking an e-commerce order processing service as an example, the target optimization index covers the order processing time, the average processing time can be controlled within a specific threshold value, such as within 2 hours, the inventory turnover rate is expected to be maintained in a trust interval through reasonable decision, the fund liquidity is ensured, for example, 3-5 times of turnover per month, and the customer return rate is reduced to a certain proportion of not more than 5%, so that the product and the service quality are reflected to meet the market demand.

And when the intelligent decision system outputs a processing decision aiming at a certain batch of E-commerce orders according to the current business configuration model, immediately starting an evaluation flow. The actual time consumption of each link from receiving to delivering of the order is monitored in real time, real-time inventory change data fed back by an inventory system and customer return information counted by an after-sales department are compared, and the real-time inventory change data and the customer return information are checked with preset target optimization indexes one by one. When the average processing time of the order reaches 3 hours and exceeds the preset standard of 2 hours, the system rapidly cuts into the depth analysis mode at the moment, and the parameter deviation item causing the delay is extracted. The investigation is that in the logistics distribution link, the distribution vehicles in a certain area are unreasonably scheduled, so that part of orders wait for the distribution time to be too long, and the logistics distribution vehicle distribution proportion is accurately locked as a key parameter deviation item.

S520, acquiring all relevant deviation factors related to the parameter deviation items, and adjusting a current service configuration model or adopting other optimization strategies based on all relevant deviation factors until the target optimization index is met.

Specifically, continuing with the previous example, after locking the parameter bias term "distribution ratio of logistics distribution vehicles", the system may develop a full search, mining all relevant bias factors closely related thereto. The relevant deviation factors influencing the vehicle distribution proportion comprise real-time fluctuation conditions of order quantities of different areas, sudden increase of the order quantities of certain hot areas during sales promotion activities, untimely follow-up of original vehicle distribution, road condition information, partial distribution routes become congested due to sudden conditions such as construction and traffic control, but are not fully considered during vehicle dispatching, vehicle loading efficiency, actual cargo carrying capacity of certain vehicle types is far lower than design capacity due to unreasonable cargo loading layout, and overall distribution efficiency is reduced.

Based on these relevant bias factors, the system rapidly formulates an optimization strategy. On one hand, the method is used for accurately adjusting a current service configuration model, a real-time order quantity monitoring module is introduced into a logistics distribution sub-model, so that the real-time order quantity monitoring module can automatically optimize the vehicle distribution proportion according to the dynamic changes of the order quantities of different areas, meanwhile, a road condition real-time updating function is embedded, map navigation data is combined, an optimal route is planned for distribution vehicles, a congestion road section is avoided, a cargo loading algorithm is optimized, a vehicle loading scheme is reasonably arranged according to the commodity types and volumes of orders, and the loading efficiency is improved. On the other hand, if the model is difficult to take effect quickly in a short period, other auxiliary optimization strategies are immediately started, such as temporarily increasing outsourcing delivery vehicles, relieving insufficient capacity in order peak period, establishing an information communication channel with traffic departments, acquiring road condition early warning in advance, planning a standby route in advance and the like.

And continuously and circularly executing the two steps, continuously evaluating a decision result according to the target optimization index, accurately extracting and overcoming the parameter deviation item and the related deviation factor thereof, flexibly adjusting the model and the application strategy until the e-commerce order processing service comprehensively meets all preset target optimization indexes, and improving the service operation efficiency.

Similarly, under other business scenes such as financial credit approval, manufacturing production scheduling and the like, the evaluation and optimization flow can be followed, and the adaptive evaluation method and adjustment strategy can be customized individually only according to the unique target optimization index, parameter characteristics and deviation factor association logic of each industry, so that the intelligent decision system can be accurately adaptive in all fields.

In a specific embodiment, as shown in fig. 5, in step S130, the service configuration model is trimmed in real time based on real-time data, to obtain the current service configuration model, which specifically includes the following steps:

s610, analyzing the complexity, real-time requirements and data volume of the real-time service data, and determining the model type and model complexity.

Specifically, the system starts the deep analysis process first in the face of real-time business data added in real time in different business scenes. The remote medical diagnosis business in the smart medical field is taken as an example for explanation. The complexity of the real-time business data is represented by information intersection of multi-dimensional information, on one hand, the real-time business data comprises various physiological monitoring data uploaded by a patient, such as continuous fluctuation data sequences of heart rate, blood pressure, blood sugar and the like, interpretation of the data needs professional medical knowledge, potential association and mutual influence exist among different indexes, and on the other hand, unstructured or semi-structured data such as medical history data, image diagnosis reports and the like of the patient are needed, and key information contained in the heterogeneous data needs to be integrated and mined.

The real-time requirement of medical data is high, and for the diagnosis of some critical diseases, delay of each second may be critical to life of a patient, and the system is required to give preliminary diagnosis advice in a very short time, such as several minutes or even tens of seconds after receiving the data, so that medical staff can take treatment measures in time.

With the popularization of wearable medical equipment and the increase of medical treatment frequency of patients, data is in mass increasing speed. For example, a large telemedicine platform may receive various types of monitoring data from tens of thousands of patients daily, accumulating up to several GB and more.

Based on an accurate parsing of these characteristics, the system determines the model type and complexity to adapt to. Aiming at the remote medical diagnosis service, an intelligent diagnosis model fused with a deep learning algorithm is selected in consideration of the complexity of data, the complex mapping relation between physiological data and diseases can be automatically learned, the medical history data is analyzed by utilizing a natural language processing technology, the model structure is required to be optimized and simplified in view of high real-time requirements, unnecessary calculation levels are reduced, rapid operation is ensured, a distributed calculation architecture support model is adopted to run in the face of large data volume, the data processing capacity is improved, the high-efficiency analysis requirement of massive information is met, and the model architecture which can meet the service requirement and has high-efficiency execution capacity is determined.

S620, introducing a model library comprising a pre-trained model template, and generating a current service configuration model according to the model type and the model complexity.

Specifically, after the analysis is completed, the system introduces a model library that includes a plurality of pre-trained model templates. The templates are constructed based on accumulated data and practical experience in different industries and various business scenes, and have wide applicability and universality.

Also taking the remote medical diagnosis service as an example, a medical diagnosis model template which is most matched with the type and complexity of the model determined before is screened from a model library. The template is pre-trained by utilizing massive medical data in the early stage, and has preliminary cognition on characteristic modes of common diseases, association of physiological index anomalies and diseases and the like.

And according to the template, carrying out targeted fine adjustment by combining the real-time data characteristics of the current remote medical service. In terms of data complexity, aiming at the special disease distribution and regional high-incidence disease characteristics of local patient groups, a feature extraction layer of the model is optimized to focus on the disease related indexes focused locally, super parameters of the model are further adjusted based on real-time requirements, such as an iterative calculation period is shortened, a memory use strategy is optimized, the model is ensured to output a diagnosis result in a specified time, and distributed calculation resources are reasonably configured in consideration of data quantity factors, so that the model is more robust and efficient in large-scale data processing.

Through the accurate adaptation and adjustment process, the current service configuration model which completely fits the current remote medical diagnosis service requirement is finally generated, and support is provided for the follow-up accurate and efficient intelligent decision.

Similarly, in other fields such as intelligent traffic control and industrial automatic production, the flow is strictly followed by the traffic data characteristics, such as road condition real-time data in intelligent traffic and vehicle running track data, equipment running parameters in industrial production and product quality monitoring data, so that the generated traffic configuration model is closely related to actual traffic, and efficient and intelligent operation management is realized.

And/or

After the intelligent decision result corresponding to the service type is generated by the decision logic module, the method specifically comprises the following steps:

s630, verifying the intelligent decision result by adopting at least one method of cross verification, A/B test or analog simulation, and obtaining a verification result.

S640, continuing to adjust the current service configuration model or introducing an adjustment data source according to the verification result.

Specifically, at least one method of cross validation, A/B test or analog simulation is adopted to strictly validate the intelligent decision result. The method comprises the steps of carrying out cross verification, carrying out multiple grouping tests, comprehensively evaluating the stability and reliability of a model, carrying out A/B test to compare the performance difference of different decision schemes in an actual service scene, screening an optimal scheme, simulating service operation by using a virtual environment through simulation, and pre-judging the policy effect in advance. After the verification result is obtained, if the decision result is found to be insufficient, the current service configuration model is continuously and finely adjusted according to the result feedback, or an adjustment data source is introduced, and the decision flow is re-optimized until the decision result reaches an effective state.

In a specific embodiment, the intelligent decision method based on service configuration further specifically includes a user interaction interface for displaying service configuration options, decision process and optimization suggestions, and allowing a user to perform configuration adjustment or result confirmation.

The system is provided with an intuitively friendly user interaction interface, clearly displays service configuration options, enables a user to conveniently adjust model parameters according to actual service demands, presents a decision process in real time, enables the user to know decision basis and logic deduction in depth, provides optimization suggestions, and assists the user in making more intelligent decisions. Meanwhile, the user is allowed to confirm the final decision result, and smoothness and acceptance of decision landing are ensured.

In a specific embodiment, the intelligent decision method based on the service configuration further specifically comprises the steps of establishing a risk early warning mechanism and carrying out real-time monitoring and early warning on potential risk points.

Specifically, a comprehensive risk early warning mechanism is established, and uninterrupted real-time monitoring is carried out on potential risk points in the service operation process. The risk, such as business process blocking risk, resource exhaustion risk, user loss risk and the like, is pre-judged in advance through comprehensive analysis of multidimensional information such as real-time business data, model running state, external environment factors and the like, and accurate early warning is timely sent out, precious time is strived for enterprises to take countermeasures in time, and steady and continuous development of the business is guaranteed.

It should be understood that the sequence number of each step in the foregoing embodiment does not mean that the execution sequence of each process should be determined by the function and the internal logic, and should not limit the implementation process of the embodiment of the present invention.

In an embodiment, an intelligent decision device based on service configuration is provided, where the intelligent decision device based on service configuration corresponds to the intelligent decision method based on service configuration in the above embodiment one by one. As shown in fig. 6, the intelligent decision device based on service configuration includes a service data acquisition module 110, a service data obtaining module 120, a service model obtaining module 130, a service data importing module 140 and a decision result generating module 150. The functional modules are described in detail as follows:

The service data acquisition module 110 is configured to acquire real-time service data in a service system, where the real-time service data includes service type, service flow information, service resource status information, and user interaction information.

The service data obtaining module 120 is configured to learn and identify noise data features by using a machine learning algorithm, dynamically update a data cleaning rule, clean real-time service data according to the updated data cleaning rule, and remove noise data and redundant data to obtain standardized service data.

The service model obtaining module 130 is configured to determine a matched historical service configuration model according to the service type, and perform real-time fine tuning on the service configuration model based on real-time data to obtain a current service configuration model.

The service data importing module 140 is configured to import the standardized service data into the current service configuration model.

The decision result generating module 150 is configured to analyze and process the standardized service data through the current service configuration model, and generate an intelligent decision result corresponding to the service type through the decision logic module.

For specific limitations of the intelligent decision device based on the service configuration, reference may be made to the above limitation of the intelligent decision method based on the service configuration, and no further description is given here. The above-mentioned intelligent decision-making device based on service configuration may be implemented in whole or in part by software, hardware or a combination thereof. The above modules may be embedded in hardware or independent of a processor in the electronic device, or may be stored in software in a memory in the electronic device, so that the processor may call and execute operations corresponding to the above modules.

In one embodiment, an electronic device is provided, which may be a server, and an internal structure thereof may be as shown in fig. 7. The electronic device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the electronic device is configured to provide computing and control capabilities. The memory of the electronic device includes a non-volatile medium, an internal memory. The non-volatile medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile media. The database of the electronic device is used for data related to the intelligent decision method based on service configuration. The network interface of the electronic device is used for communicating with an external terminal through a network connection. The computer program, when executed by a processor, implements an intelligent decision method based on service configuration.

In an embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor executes the computer program to implement the intelligent decision method based on service configuration according to the above embodiment, for example, S110 to S150 shown in fig. 2. Or the processor, when executing the computer program, implements the functions of the modules/units of the intelligent decision apparatus based on service configuration in the above embodiment, such as the functions of the modules 110 to 150 shown in fig. 6. To avoid repetition, no further description is provided here.

In one embodiment, a computer readable medium is provided, on which a computer program is stored, which when executed by a processor implements the intelligent decision method based on service configuration of the above embodiment, for example, S110 to S150 shown in fig. 2. Or the computer program, when executed by a processor, implements the functions of the modules/units in the intelligent decision-making device based on service configuration in the above-described device embodiment, such as the functions of the modules 110 to 150 shown in fig. 6. To avoid repetition, no further description is provided here.

Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable medium that when executed comprises the steps of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments of the application may include non-volatile and/or volatile memory. The non-volatile memory may include a read only memory (RM), a Programmable RM (PRM), an Electrically Programmable RM (EPRM), an Electrically Erasable Programmable RM (EEPRM), or a flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (SYNCHLINK) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.

It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions.

The foregoing embodiments are merely for illustrating the technical solution of the present invention, but not for limiting the same, and although the present invention has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solution described in the foregoing embodiments may be modified or substituted for some of the technical features thereof, and that these modifications or substitutions should not depart from the spirit and scope of the technical solution of the embodiments of the present invention and should be included in the protection scope of the present invention.

Claims

1. An intelligent decision-making method based on business configuration, characterized by comprising:

Collecting real-time business data in the business system, the real-time business data includes business type, business process information, business resource status information and user interaction information;

Using a machine learning algorithm to learn and identify the characteristics of noise data, dynamically updating data cleaning rules, cleaning the real-time business data according to the updated data cleaning rules, removing noise data and redundant data, and obtaining standardized business data;

According to the business type, a matching historical business configuration model is determined, and the business configuration model is fine-tuned in real time based on real-time data to obtain a current business configuration model;

Importing the standardized business data into the current business configuration model;

The standardized business data is analyzed and processed through the current business configuration model, and an intelligent decision result corresponding to the business type is generated through a decision logic module.

2. The intelligent decision-making method based on business configuration according to claim 1, characterized in that the real-time business data collected in the business system includes:

Adopting a distributed data collection architecture, setting up data collection probes at multiple key nodes of the business system, and aggregating the collected data into a data buffer pool through asynchronous transmission;

The method according to claim 1, the intelligent decision-making method based on business configuration according to claim 1, characterized in that it also includes:

Regularly update and maintain the business configuration model, including:

At preset time intervals, new historical business data is collected to update and improve the business rules in the business rule library, supplement new business rules or amend old rules;

Recalibrate the key parameters and their threshold ranges in the business parameter library according to business development needs and industry dynamic adjustments;

The decision logic module is optimized and trained using a reinforcement learning algorithm.

3. The intelligent decision-making method based on business configuration according to claim 1, characterized in that before determining the matching historical business configuration model, it also includes:

An initial business configuration model is constructed based on historical business data and industry standard business processes, and is used to generate a historical business configuration model based on real-time data and the initial business configuration model, wherein the initial business configuration model includes: a business rule library, a business parameter library, and a decision logic module, wherein the business rule library stores the operation rules of various types of businesses, the business parameter library stores the key parameters in the business process and their threshold ranges, and the decision logic module generates decision recommendations based on the business rules and parameter threshold ranges.

4. The intelligent decision-making method based on business configuration according to claim 1, characterized in that the real-time business data also includes target optimization indicators;

After the intelligent decision result corresponding to the business type is generated by the decision logic module, the method further includes:

Evaluate the intelligent decision-making result according to the target optimization index, and if it does not meet the preset standard, extract the parameter deviation item;

All relevant deviation factors related to the parameter deviation item are obtained, and based on all the relevant deviation factors, the current business configuration model is adjusted or other optimization strategies are adopted until the target optimization index is met.

5. The intelligent decision-making method based on business configuration according to claim 1, characterized in that the real-time fine-tuning of the business configuration model based on real-time data to obtain the current business configuration model comprises:

Analyze the complexity, real-time requirements and data volume of the real-time business data to determine the model type and model complexity;

Introducing a model library including pre-trained model templates, and generating the current business configuration model according to the model type and the model complexity;

and/or

Verify the intelligent decision-making result by using at least one of cross-validation, A/B testing or simulation, and obtain the verification result;

The current business configuration model continues to be adjusted or an adjustment data source is introduced according to the verification result.

6. According to claim 1, the intelligent decision-making method based on business configuration is characterized in that it also includes a user interaction interface for displaying business configuration options, decision-making process and optimization suggestions, allowing users to make configuration adjustments or confirm results.

7. According to claim 1, the intelligent decision-making method based on business configuration is characterized in that it also includes establishing a risk warning mechanism to conduct real-time monitoring and warning of potential risk points.

8. An intelligent decision-making device based on business configuration, characterized by comprising:

A business data collection module is used to collect real-time business data in the business system, wherein the real-time business data includes business type, business process information, business resource status information and user interaction information;

A business data acquisition module is used to learn and identify noise data features using a machine learning algorithm, dynamically update data cleaning rules, clean the real-time business data according to the updated data cleaning rules, remove noise data and redundant data, and obtain standardized business data;

A business model acquisition module is used to determine a matching historical business configuration model according to the business type, and to fine-tune the business configuration model in real time based on real-time data to obtain a current business configuration model;

A business data import module, used to import the standardized business data into the current business configuration model;

The decision result generation module is used to analyze and process the standardized business data through the current business configuration model, and generate an intelligent decision result corresponding to the business type through a decision logic module.

9. An electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the intelligent decision-making method based on business configuration as described in any one of claims 1 to 7 is implemented.

10. A computer-readable medium storing a computer program, wherein the computer program, when executed by a processor, implements the intelligent decision-making method based on business configuration as claimed in any one of claims 1 to 7.