CN120852064A - Quality control method and device based on insurance claim, electronic equipment and storage medium - Google Patents
Quality control method and device based on insurance claim, electronic equipment and storage mediumInfo
- Publication number
- CN120852064A CN120852064A CN202510962763.2A CN202510962763A CN120852064A CN 120852064 A CN120852064 A CN 120852064A CN 202510962763 A CN202510962763 A CN 202510962763A CN 120852064 A CN120852064 A CN 120852064A
- Authority
- CN
- China
- Prior art keywords
- feature
- insurance
- materials
- quality control
- quality
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Abstract
本申请实施例提供了一种基于保险理赔的质控方法、装置、电子设备及存储介质，属于计算机技术领域，适用于金融科技领域。该方法包括：获取保险理赔材料以及保险理赔材料的理赔数量；根据预先获取的质控并发数量和理赔数量，确定至少两个序号区间；按照序号区间之间的区间排列顺序获取序号区间，得到目标区间，目标区间包括区间右端点；通过异步线程对保险理赔材料进行筛选，得到具有材料序号和目标材料类别的参考理赔材料，材料序号位于目标区间；按照材料序号，通过目标材料类别和异步线程调用质控模型依次对参考理赔材料进行质量检测，直至材料序号等于区间右端点，得到质量检测的质控结果，能够提高保险理赔材料的质控效率。
The embodiment of the present application provides a quality control method, device, electronic device and storage medium based on insurance claims, which belongs to the field of computer technology and is applicable to the field of financial technology. The method includes: obtaining insurance claim materials and the number of claims of insurance claim materials; determining at least two serial number intervals according to the pre-acquired quality control concurrent number and the number of claims; obtaining the serial number intervals according to the interval arrangement order between the serial number intervals to obtain the target interval, and the target interval includes the right endpoint of the interval; screening the insurance claim materials through an asynchronous thread to obtain a reference claim material with a material serial number and a target material category, and the material serial number is located in the target interval; according to the material serial number, the reference claim material is sequentially quality-checked through the target material category and the asynchronous thread calling the quality control model until the material serial number is equal to the right endpoint of the interval, and the quality control result of the quality check is obtained, which can improve the quality control efficiency of the insurance claim materials.
Description
Technical Field
The application relates to the technical field of computers, and is applied to the field of financial science and technology, in particular to a quality control method, device, electronic equipment and storage medium based on insurance claim settlement.
Background
When a user applies for insurance claims to an insurance institution, the insurance claim materials such as identity card portrait pages, invoices and the like are uploaded through a service interface provided by the insurance institution. In order to ensure the accuracy of the insurance claim material, the quality control needs to be performed on the insurance claim material, such as judging whether the identity card portrait page is clear or not. Along with the expansion of insurance business scale, insurance institutions receive a large amount of insurance claim materials, and traditional mode adopts manual auditing mode to carry out quality control to insurance claim materials for the quality control period is longer, leads to the quality control efficiency of insurance claim materials lower.
Disclosure of Invention
The embodiment of the application mainly aims to provide a quality control method, a quality control device, electronic equipment and a storage medium based on insurance claim settlement, and aims to improve the quality control efficiency of insurance claim settlement materials.
To achieve the above object, a first aspect of an embodiment of the present application provides a quality control method based on insurance claim, the method including:
Acquiring an insurance claim material and the claim number of the insurance claim material;
Determining at least two sequence number intervals according to the pre-acquired quality control concurrency quantity and the claim settlement quantity;
Acquiring the sequence number intervals according to the interval arrangement sequence among the sequence number intervals to obtain a target interval, wherein the target interval comprises an interval right endpoint;
Screening the insurance claim materials through a preset asynchronous thread to obtain reference claim materials with material serial numbers and target material categories, wherein the material serial numbers are positioned in the target intervals;
And according to the material serial number, sequentially carrying out quality detection on the reference claim material through the target material class and the asynchronous thread call quality control model until the material serial number is equal to the right end point of the interval to obtain a quality control result of quality detection, wherein the quality control result comprises that the quality detection is passed or the quality detection is failed.
In some embodiments, the obtaining the insurance claim material and the claim amount of the insurance claim material includes:
acquiring original claim material;
classifying the materials of the original claim materials to obtain predicted material types;
Performing preliminary detection on the original claim material according to the predicted material category to obtain a material quality category;
Screening the original claim settlement materials according to the predicted material category and the material quality category to obtain the insurance claim settlement materials;
the claim amount of the insurance claim material is obtained.
In some embodiments, the classifying the material of the original claim material to obtain a predicted material class includes:
Extracting features of the original claim material to obtain first-mode claim feature and second-mode claim feature;
performing cross attention calculation on the first-modality claim feature and the second-modality claim feature to obtain an attention feature;
Performing feature fusion on the first-mode claim feature, the second-mode claim feature and the attention feature to obtain a reference claim feature;
and classifying the materials of the original claim according to the reference claim characteristics to obtain the predicted material category.
In some embodiments, the cross-attention computing of the first modality claim feature and the second modality claim feature to obtain an attention feature includes:
performing feature mapping on the first modal claim feature to obtain a query feature;
Performing feature mapping on the second-modality claim feature to obtain a key feature and a value feature;
Calculating an attention score from the query feature and the key feature;
And carrying out weighted summation on the value characteristics according to the attention score to obtain the attention characteristics.
In some embodiments, the feature fusion of the first modality claim feature, the second modality claim feature, and the attention feature, to obtain a reference claim feature, includes:
Calculating a gating weight value according to the attention characteristic;
Performing feature fusion on the first modal claim feature and the attention feature according to the gating weight value to obtain a first fusion feature;
performing feature fusion on the second modal claim feature and the attention feature according to the gating weight value to obtain a second fusion feature;
And carrying out feature fusion on the first fusion feature and the second fusion feature to obtain the reference claim feature.
In some embodiments, the feature fusing the first fused feature and the second fused feature to obtain the reference claim feature includes:
calculating the similarity between the first fusion feature and the second fusion feature;
performing principal component analysis on the first fusion feature according to the similarity to obtain a first key feature;
Performing principal component analysis on the second fusion feature to obtain a second key feature;
And carrying out feature fusion on the first key features and the second key features to obtain the reference claim features.
In some embodiments, the preliminary detection of the original claim material according to the predicted material class, to obtain a material quality class, includes:
performing optical character recognition on the original claim material to obtain a character sequence;
Carrying out keyword recognition on the character sequence to obtain a target keyword;
calculating a quality score of the original claim material according to the predicted material category and the target keyword;
and classifying the quality of the original claim material according to the quality score to obtain the quality class of the material.
To achieve the above object, a second aspect of the embodiments of the present application provides a quality control device based on insurance claims, the device including:
The material acquisition module is used for acquiring insurance claim materials and the claim number of the insurance claim materials;
The interval determining module is used for determining at least two sequence number intervals according to the quality control concurrency quantity and the claim settlement quantity which are acquired in advance;
The interval acquisition module is used for acquiring the sequence number intervals according to the interval arrangement sequence among the sequence number intervals to obtain a target interval, wherein the target interval comprises an interval right endpoint;
the screening module is used for screening the insurance claim materials through a preset asynchronous thread to obtain reference claim materials with material serial numbers and target material categories, wherein the material serial numbers are located in the target intervals;
And the quality control module is used for sequentially carrying out quality detection on the reference claim material through the target material class and the asynchronous thread according to the material serial number until the material serial number is equal to the right end point of the interval to obtain a quality control result of quality detection, wherein the quality control result comprises that the quality detection is passed or the quality detection is failed.
To achieve the above object, a third aspect of the embodiments of the present application proposes an electronic device, including a memory storing a computer program and a processor implementing the method according to the first aspect when the processor executes the computer program.
To achieve the above object, a fourth aspect of the embodiments of the present application proposes a computer-readable storage medium storing a computer program which, when executed by a processor, implements the method of the first aspect.
The quality control method based on the insurance claims, the quality control device based on the insurance claims, the electronic equipment and the computer readable storage medium provided by the embodiment of the application acquire the insurance claims materials and the claims amount of the insurance claims materials so as to control the quality of the insurance claims materials according to the claims amount. When the number of the claims is large, in order to improve the quality control efficiency, at least two sequence number intervals are determined according to the obtained quality control concurrency number and the obtained claim number in advance, so that insurance claim materials are distributed to different sequence number intervals, the workload can be distributed more reasonably, and the resource utilization efficiency is improved. And acquiring sequence number intervals according to the interval arrangement sequence among the sequence number intervals to obtain a target interval so as to control the quality of insurance claim materials positioned in the target interval. In order to improve quality control efficiency, an asynchronous thread is started, and insurance claim materials are screened through a preset asynchronous thread to obtain reference claim materials with material serial numbers and target material types, wherein the material serial numbers are located in target intervals, so that the insurance claim materials which need quality control of the asynchronous thread are determined. In order to realize an automatic quality control flow, quality detection is sequentially carried out on the reference claim material through the target material class and the asynchronous thread according to the material serial number until the material serial number is equal to the right end point of the interval, so that a quality control result of quality detection of the insurance claim material with the material serial number in the target interval is obtained, and compared with an artificial quality control mode, the quality control efficiency is greatly improved.
Drawings
FIG. 1 is a flow chart of a quality control method based on insurance claims provided by an embodiment of the present application;
Fig. 2 is a flowchart of step S110 in fig. 1;
Fig. 3 is a flowchart of step S220 in fig. 2;
Fig. 4 is a flowchart of step S320 in fig. 3;
fig. 5 is a flowchart of step S330 in fig. 3;
fig. 6 is a flowchart of step S540 in fig. 5;
Fig. 7 is a flowchart of step S230 in fig. 2;
Fig. 8 is a schematic structural diagram of a quality control device based on insurance claims according to an embodiment of the present application;
fig. 9 is a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
It should be noted that although functional block division is performed in a device diagram and a logic sequence is shown in a flowchart, in some cases, the steps shown or described may be performed in a different order than the block division in the device, or in the flowchart. The terms first, second and the like in the description and in the claims and in the above-described figures, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of the application only and is not intended to be limiting of the application.
When a user applies for insurance claims to an insurance institution, the insurance claim materials such as identity card portrait pages, invoices and the like are uploaded through a service interface provided by the insurance institution. In order to ensure the accuracy of the insurance claim material, the quality control needs to be performed on the insurance claim material, such as judging whether the identity card portrait page is clear or not. Along with the expansion of insurance business scale, insurance institutions receive a large amount of insurance claim materials, and traditional mode adopts manual auditing mode to carry out quality control to insurance claim materials for the quality control period is longer, leads to the quality control efficiency of insurance claim materials lower.
Based on the above, the embodiment of the application provides a quality control method based on insurance claims, a quality control device based on insurance claims, electronic equipment and a computer readable storage medium, aiming at improving the quality control efficiency of insurance claims materials.
The quality control method based on the insurance claim, the quality control device based on the insurance claim, the electronic equipment and the computer readable storage medium provided by the embodiment of the application are specifically described through the following embodiments, and the quality control method based on the insurance claim in the embodiment of the application is described first.
The embodiment of the application provides a quality control method based on insurance claim settlement, which relates to the technical field of computers. The quality control method based on insurance claim provided by the embodiment of the application can be applied to a terminal, a server side and software running in the terminal or the server side. In some embodiments, the terminal may be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc., the server may be configured as an independent physical server, may be configured as a server cluster or a distributed system formed by a plurality of physical servers, and may be configured as a cloud server for providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligent platforms, and the software may be an application for implementing a quality control method based on insurance claims, but is not limited to the above form.
The application is operational with numerous general purpose or special purpose computer system environments or configurations. Such as a personal computer, a server computer, a hand-held or portable device, a tablet device, a multiprocessor system, a microprocessor-based system, a set top box, a programmable consumer electronics, a network PC, a minicomputer, a mainframe computer, a distributed computing environment that includes any of the above systems or devices, and the like. The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Fig. 1 is an optional flowchart of a quality control method based on insurance claims according to an embodiment of the present application, where the method in fig. 1 may include, but is not limited to, steps S110 to S150.
Step S110, acquiring insurance claim materials and the claim number of the insurance claim materials;
step S120, determining at least two sequence number intervals according to the pre-acquired quality control concurrency quantity and claim settlement quantity;
Step S130, obtaining a sequence number interval according to an interval arrangement sequence among the sequence number intervals to obtain a target interval, wherein the target interval comprises an interval right endpoint;
step S140, screening insurance claim materials through a preset asynchronous thread to obtain reference claim materials with material serial numbers and target material categories, wherein the material serial numbers are located in target intervals;
And step S150, sequentially carrying out quality detection on the reference claim material through the target material class and the asynchronous thread according to the material serial number, and obtaining a quality control result of the quality detection until the material serial number is equal to the right end point of the interval, wherein the quality control result comprises that the quality detection is passed or the quality detection is failed.
In the steps S110 to S150 shown in the embodiment of the application, through automatic quality control of the insurance claim material, the whole quality control process does not need manual intervention, and the quality control efficiency of the insurance claim is improved.
Referring to fig. 2, in some embodiments, step S110 may include, but is not limited to, steps S210 to S250:
Step S210, acquiring original claim material;
Step S220, classifying the materials of the original claim materials to obtain predicted material types;
Step S230, carrying out preliminary detection on the original claim material according to the predicted material category to obtain the material quality category;
Step S240, screening the original claim settlement materials according to the predicted material types and the material quality types to obtain insurance claim settlement materials;
step S250, obtaining the claim number of the insurance claim material.
In step S210 of some embodiments, when the user needs to apply for an insurance claim, various documents are uploaded through a business service interface provided by the insurance agency to obtain the original claim material. The original claim material is related to the insurance type of the claim to be settled, for example, the insurance type is health insurance, and the original claim material comprises identity cards, medical diagnosis results, medical records, medical expense invoices and the like, and the identity cards can be identity cards and household books. For another example, the insurance type is car insurance, and the original claim material comprises identification, accident scene photo, vehicle maintenance invoice, accident responsibility identification and the like.
In step S220 of some embodiments, the original claim material includes various proof materials, and quality control standards of the different proof materials are different, for example, the front side of the identification card needs to include key information such as name, identification card number, etc., and the invoice needs to include information such as bill, amount, etc. In order to improve the quality control accuracy of insurance claims, the material type of the proof material needs to be determined so that the proof material is subsequently quality controlled according to the material type. Specifically, the material classification is carried out on the original claim material through a classification model, and the original claim material is distributed to a predefined label to obtain a predicted material class. The classification model may be a convolutional neural network, a cyclic neural network, a transducer model, or the like. Predefined tags include identity front, identity card back, account book, invoice, document, and others. The labels of the medical record, the medical diagnosis result and the accident responsibility identification are all documents. The original claim material includes a plurality of proof materials, and the predicted material class includes a material type for each proof material. The predicted material categories include at least one of identity face-to-face, identity card face-to-face, account book, invoice, document, and others.
In step S230 of some embodiments, the original claim material includes a plurality of proof materials, and the predicted material class includes a material type for each proof material. And for each proving material, performing preliminary detection on the proving material according to the material type to obtain the detection result of the proving material. If the detection result of each proving material is that the proving material passes the detection, determining that the material quality class of the original claim material is up to the quality standard. If the detection result of the proof material is that the proof material detection fails, determining that the material quality class of the original claim material is that the quality does not reach the standard.
In step S240 of some embodiments, the preset material class is the material type of the various proof materials that should be uploaded for the insurance type to be claiming. And comparing the predicted material category with the preset material category, and taking the original claim settlement material as the insurance claim settlement material if the predicted material category comprises all material types of the preset material category and the material quality category is up to the quality standard. If the predicted material class lacks the material type of the preset material class, the original claim material uploaded by the user is indicated to lack a certain proof material, or the predicted material class is of unqualified quality, prompt information is sent to the user through the business service interface, so that the user can upload the missing proof material again or detect the unqualified proof material according to the prompt information, a subsequent quality control process is not needed, and the quality control efficiency of insurance claim is improved.
In step S250 of some embodiments, the amount of insurance claim material is obtained, resulting in the amount of claim to be resolved. It should be noted that, a plurality of proof materials for the same insurance accident are one insurance claim material, and the number of claims is 1.
Through the steps S210 to S250, the insurance claim materials which are qualified in preliminary detection can be screened from the mass claim materials, the claim number of the insurance claim materials with subsequent quality control is reduced, and the quality control efficiency of the insurance claim is improved.
Referring to fig. 3, in some embodiments, step S220 may include, but is not limited to, steps S310 to S340:
Step S310, extracting features of the original claim material to obtain first-mode claim feature and second-mode claim feature;
Step S320, cross attention calculation is carried out on the first mode claim feature and the second mode claim feature to obtain attention features;
step S330, carrying out feature fusion on the first-mode claim feature, the second-mode claim feature and the attention feature to obtain a reference claim feature;
And step S340, classifying the materials of the original claim according to the reference claim characteristics to obtain the predicted material category.
In step S310 of some embodiments, the classification model includes a plurality of feature extraction layers and classification layers that are serially connected in sequence. And carrying out feature extraction on the original claim material through the first feature extraction layer to obtain the claim feature output by the first feature extraction layer. And inputting the claim features output by the last feature extraction layer into the current feature extraction layer for feature extraction to obtain the claim features output by the current feature extraction layer, and repeating the steps until the claim features output by the last feature extraction layer are obtained. And taking the claim-settling characteristics output by any two characteristic extraction layers as a first-mode claim-settling characteristic and a second-mode claim-settling characteristic. The first mode claim feature and the second mode claim feature are the claim features of different modes extracted by different feature extraction layers.
In step S320 of some embodiments, in order to capture the complex dependency between the first modality claim feature and the second modality claim feature, to better understand the feature interactions between the different feature extraction layers, the cross attention computation is performed on the first modality claim feature and the second modality claim feature of each two feature extraction layers, resulting in a plurality of attention features.
In step S330 of some embodiments, in order to reduce information loss, enhance feature expression capability and generalization capability of the classification model, feature stitching or feature addition is performed on the first modality claim feature, the second modality claim feature, and the plurality of attention features, to obtain the reference claim feature.
In step S340 of some embodiments, the reference claim signature is input to a classification layer to classify the original claim material into a predicted material class. The classification layer can be a full connection layer, a multi-layer sensor, etc.
Through the steps S310 to S340, the predicted material category can be obtained, so as to determine whether the original claim material lacks the proof material according to the predicted material category, thereby realizing the preliminary detection and screening of the original claim material, and further improving the quality control efficiency of the insurance claim.
Referring to fig. 4, in some embodiments, step S320 may include, but is not limited to, steps S410 to S440:
step S410, performing feature mapping on the first mode claim feature to obtain a query feature;
Step S420, performing feature mapping on the second mode claim feature to obtain a key feature and a value feature;
step S430, calculating attention scores according to the query characteristics and the key characteristics;
in step S440, the value features are weighted and summed according to the attention score to obtain the attention feature.
In step S410 of some embodiments, the query weight is multiplied by the first modality claim feature to obtain a query feature.
In step S420 of some embodiments, the key weight and the second modality claim feature are multiplied to obtain a key feature. And multiplying the value weight and the second mode claim feature to obtain the value feature.
In step S430 of some embodiments, dot product computation is performed on the query feature and the key feature, resulting in a similarity between the query feature and the key feature. And determining a scaling factor according to the feature dimension of the key feature, and scaling the similarity according to the scaling factor. And normalizing the similarity after the scaling treatment through a softmax function to obtain the attention score. The calculation formula of the attention score is expressed as:
Wherein Q represents a query feature, K represents a key feature, T represents a transpose operation, d represents a feature dimension of the key feature; Representing the scaling factor.
In step S440 of some embodiments, the attention score and the value characteristic are multiplied to obtain an attention characteristic. If there are multiple attention heads, the attention score output by each attention head can be multiplied by the value characteristic, and the calculation results obtained by the multiplication are added to obtain the attention characteristic.
The two feature extraction layers are mutually influenced, not unidirectionally influenced, and in order to ensure the extraction accuracy of the attention features, the query weight and the second-mode claim feature can be multiplied to obtain the intermediate query feature. And multiplying the key weight and the first mode claim feature to obtain an intermediate key feature, and multiplying the value weight and the first mode claim feature to obtain an intermediate value feature. An intermediate attention score is calculated based on the intermediate query feature and the intermediate key feature. And weighting and summing the intermediate value features according to the intermediate attention score to obtain the intermediate attention feature. And carrying out average value calculation on the attention characteristic and the middle attention characteristic to obtain a final attention characteristic.
If the attention feature is denoted as h1, the intermediate attention feature is denoted as h2, and the final attention feature is denoted as (h1+h2)/2.
Through the steps S410 to S440, complex dependency relationships between different modal features can be captured, and a more comprehensive feature representation is provided, so that the accuracy of classifying the claim materials is improved.
Referring to fig. 5, in some embodiments, step S330 may include, but is not limited to, steps S510 to S540:
step S510, calculating a gating weight value according to the attention characteristic;
step S520, carrying out feature fusion on the first modal claim feature and the attention feature according to the gating weight value to obtain a first fusion feature;
Step S530, carrying out feature fusion on the second mode claim feature and the attention feature according to the gating weight value to obtain a second fusion feature;
And S540, carrying out feature fusion on the first fusion feature and the second fusion feature to obtain the reference claim feature.
In step S510 of some embodiments, for each attention feature, feature mapping is performed on the attention feature to obtain a gating weight value, where the gating weight value ranges between [0,1 ]. The retention or inhibition of feature dimensions can be dynamically controlled through the gating weight value, so that the fine granularity feature selection capability of the classification model is enhanced. The calculation process of the feature map is expressed as:
g=σ(WH+b),
wherein g is a gating weight value, sigma represents an activation function, which can be a sigmoid function, W and b are projection weights and offsets respectively, and H is a attention feature.
In step S520 of some embodiments, in order to achieve effective feature fusion without increasing computational complexity, feature addition is performed on the first-modality claim feature and the attention feature to obtain the first-modality feature, so that context information introduced by the attention feature is fused into the first-modality claim feature, thereby enhancing feature expression capability, enabling the classification model to better capture global and local information, and improving understanding capability of complex insurance claim materials. Multiplying the gating weight value by the first modal feature to dynamically adjust the feature importance of the first modal feature according to the gating weight value to obtain a first fusion feature.
In step S530 of some embodiments, the second modality claim feature and the attention feature are feature added to incorporate the contextual information introduced by the attention feature into the second modality claim feature, resulting in a second modality feature. Multiplying the gating weight value by the second modal feature to dynamically adjust the feature importance of the second modal feature according to the gating weight value to obtain a second fusion feature.
In step S540 of some embodiments, the first fused feature and the second fused feature calculated based on each attention feature are feature spliced or feature added to obtain a reference claim feature.
The steps S510 to S540 can implement fine granularity selection of features through a gating mechanism, so that unnecessary feature computation is reduced. By introducing the context information provided by the attention mechanism, the feature expression capability can be enhanced, and the adaptability and generalization capability of the model are improved.
Referring to fig. 6, in some embodiments, step 540 may include, but is not limited to, steps S610 through S640:
step S610, calculating the similarity between the first fusion feature and the second fusion feature;
step S620, carrying out principal component analysis on the first fusion feature according to the similarity to obtain a first key feature;
step S630, carrying out principal component analysis on the second fusion feature to obtain a second key feature;
And step S640, carrying out feature fusion on the first key features and the second key features to obtain reference claim settlement features.
In step S610 of some embodiments, a euclidean distance between the first fused feature and the second fused feature is calculated, taking the inverse of the euclidean distance as the similarity. The greater the Euclidean distance, the less similarity between the first fused feature and the second fused feature.
In step S620 of some embodiments, if the similarity is greater than or equal to the preset similarity threshold, which indicates that the first fusion feature and the second fusion feature are relatively similar, and the feature redundancy is relatively high, the principal component analysis (PRINCIPAL COMPONENT ANALYSIS, PCA) algorithm is used to perform principal component analysis on the first fusion feature, identify the main feature in the first fusion feature, and obtain the first key feature. The preset similarity threshold is a preset threshold for judging whether the features are similar or not, and can be set by self according to actual conditions, such as 0.8. If the similarity is smaller than a preset similarity threshold, the fact that the similarity between the first fusion feature and the second fusion feature is smaller is indicated, and feature fusion is conducted on the first fusion feature and the second fusion feature, so that the reference claim feature is obtained.
Specifically, the first fusion feature is divided into a plurality of row vectors or column vectors, and a plurality of vectors are obtained. And carrying out standardization processing on each vector to ensure that the mean value of each vector is 0 and the variance is 1, thereby obtaining a plurality of basic vectors. And constructing a matrix according to the positions of the basis vectors in the first fusion characteristic and the plurality of basis vectors to obtain a basis matrix. And calculating a covariance matrix of the base matrix, and solving eigenvalues and eigenvectors of the covariance matrix. And selecting the feature vector with the largest feature value to obtain a projection matrix. And multiplying the projection matrix by the first fusion feature to obtain a first key feature.
In step S630 of some embodiments, referring to step S620, principal component analysis is performed on the second fusion feature by using a principal component analysis algorithm, and the principal feature of the second fusion feature is extracted, so as to obtain a second key feature.
In step S640 of some embodiments, feature stitching or feature addition is performed on the first key feature and the second key feature calculated based on each attention feature, to obtain a reference claim feature.
Through the steps S610 to S640, the key claim settlement features can be identified, and the complexity of feature calculation is reduced, so that the classification efficiency of the insurance claim settlement materials is improved.
Referring to fig. 7, in some embodiments, step S230 may include, but is not limited to, steps S710 to S740:
step S710, performing optical character recognition on the original claim material to obtain a character sequence;
step S720, carrying out keyword recognition on the character sequence to obtain a target keyword;
Step S730, calculating the quality score of the original claim material according to the predicted material category and the target keyword;
And step S740, classifying the quality of the original claim material according to the quality scores to obtain the quality class of the material.
In some embodiments, in step S710, the original claim material is optically character-identified by an optical character recognition (Optical Character Recognition, OCR) engine, the text content of the original claim material is extracted, and the text content is converted into a text format that can be processed by a computer, resulting in a sequence of characters.
In step S720 of some embodiments, keyword recognition is performed on the character sequence by a pre-trained language model such as BERT, and keywords in the character sequence are extracted to obtain target keywords.
In step S730 of some embodiments, the predicted material class includes a material type for each proof material, and a class weight for the material type is obtained. The class weight is used for measuring the importance degree of the material type on the quality score, the larger the class weight is, the larger the importance degree is, and the range of the class weight is between [0,1 ]. And obtaining a reference keyword corresponding to the material type and a keyword weight of the reference keyword, wherein the keyword weight is used for measuring the importance degree of the keyword on the quality score, the keyword weight is in direct proportion to the importance degree, and the value range of the keyword weight is between [0 and 1 ]. And comparing the target keyword with the reference keyword, and multiplying the category weight and the keyword weight if the target keyword is the same as the reference keyword to obtain the keyword score of the target keyword. If the target keyword is different from the reference keyword, the keyword score of the target keyword is 0. And adding the keyword scores of the target keywords to obtain the quality scores of the original claim material.
In step S740 of some embodiments, if the quality score is greater than or equal to the preset score threshold, determining that the material quality class of the original claim material is quality-compliant. And if the quality score is smaller than the preset score threshold, determining that the material quality class of the original claim material is not up to the standard. The preset scoring threshold is a threshold set for judging whether the quality of the original claim material meets the standard or not, and can be set automatically according to actual conditions.
Through the steps S710 to S740, the quality of the claim material can be detected preliminarily, and the subsequent quality control process is not required for the claim material with the quality which does not reach the standard, so that the quality control efficiency of insurance claim is improved.
In step S120 of some embodiments, the insurance claim material has a material serial number, and the insurance claim material may be ordered according to an uploading time sequence of the insurance claim material or other criteria to obtain the material serial number. The earlier the upload time, the smaller the material number of the insurance claim material. The quality control concurrency quantity is the quantity of insurance claim materials to be processed in one batch, and the quality control concurrency quantity can be configured and adjusted. Dividing the number of the claims and the quality control concurrency number, and upwardly rounding the quotient obtained by dividing to obtain the number of the sections, determining a serial number section according to the number of the claims, the quality control concurrency number and the number of the sections, wherein the section length of the serial number section is the quality control concurrency number or remainder obtained by dividing calculation, and the section length is the number of the material serial numbers in the serial number section. For example, the number of claims is 23, the number of concurrency of quality control is 7, and the number of intervals is 4. And determining sequence number intervals as [1,7], [8,14], [15,21] and [22,23] according to the number of claims 23, the number of intervals 4 and the quality control concurrency number 7.
In step S130 of some embodiments, the sequence number intervals include a left end point and a right end point, and the sequence number intervals may be sorted from the left end point or the right end point from small to large to obtain an interval arrangement sequence. And acquiring a sequence number interval according to the interval arrangement sequence to obtain a target interval, wherein the target interval comprises an interval right endpoint.
In step S140 of some embodiments, an asynchronous thread is started, and insurance claim materials are screened by the asynchronous thread, and insurance claim materials with material serial numbers located in a target interval are selected to obtain reference claim materials with target material categories. The target material class is the predicted material class obtained in step S220.
In step S150 of some embodiments, according to the material serial number, the asynchronous thread sequentially calls a quality control model according to each material type in the target material class, performs quality detection on the reference claim material through the called quality control model until the material serial number is equal to the right end point of the interval, triggers a callback function to send a quality control result of quality detection corresponding to the target interval to the main thread, writes the quality control result into a log, and continues to acquire the next target interval with reference to steps S130 to S150, and performs quality detection on the next target interval through the asynchronous thread until the quality control result of the last target interval is obtained. The quality control result includes a quality detection pass or a quality detection fail. And each material type corresponds to one quality control model, and the quality control model of the corresponding material type can be called in parallel according to each material type in the target material type so as to improve the quality control efficiency of the reference claim material. In the quality detection process, the main thread cannot be blocked, when the time consumption of an asynchronous task is high, the main thread can execute other business operations, after the asynchronous task is completed, the asynchronous thread can call back a result to the main thread, the main thread does not need to wait for the asynchronous thread to return a quality detection result, and the quality control efficiency of insurance claims is greatly improved.
The embodiment of the application provides a quality control method based on insurance claim settlement, which comprises the steps that a main thread reads insurance claim settlement materials from a test path and determines the claim settlement quantity x of the insurance claim settlement materials. And determining a sequence number range as [1, x ] according to the claim number x, determining a sequence number interval as [ start_index, end_index ], wherein the start_index and the end_index are respectively the left end point and the right end point of the sequence number interval, the initial value of the start_index is 1, and the end_index=min { start_index+n-1, x }, according to the quality control concurrency number n and the sequence number range as [1, x ]. The asynchronous thread sequentially carries out quality detection on the insurance claim materials with the serial numbers in the serial number intervals, and after the quality detection on the insurance claim materials with the serial numbers equal to end_index is completed, the asynchronous thread triggers a callback function to callback the quality control result corresponding to the serial number intervals to the main thread. The main thread sets the start_index as the start_index=start_index+n, circularly initiates the asynchronous n concurrency test of the next batch, sets the end_index as the minimum value in the current start_index+n-1 and x, and determines the current sequence number interval. And detecting the quality of the insurance claim material with the serial number in the current serial number interval through an asynchronous thread until the end_index is larger than x, and ending the quality control process.
The main thread can not be blocked by calling the asynchronous thread through the main thread, other business operations can be executed at the same time, the result can be called back to the main thread after the asynchronous task is completed, when the result is called back, the quality control result is written into the log file, the next cyclic asynchronous call can be initiated according to the sequence number, and the main thread does not need to wait. After the current batch asynchronous concurrency is completed, the callback node initiates the next batch asynchronous concurrency, and the execution times of the cyclic control codes are reduced and the quality control efficiency of insurance claims is improved in comparison with the serial test.
Referring to fig. 8, an embodiment of the present application further provides a quality control device based on insurance claim, which can implement the quality control method based on insurance claim, where the quality control device based on insurance claim includes:
a material acquiring module 810, configured to acquire an insurance claim material and a claim number of the insurance claim material;
the interval determining module 820 is configured to determine at least two sequence number intervals according to the pre-acquired quality control concurrency number and the claim settlement number;
the interval acquisition module 830 is configured to acquire a sequence number interval according to an interval arrangement sequence between sequence number intervals, so as to obtain a target interval, where the target interval includes an interval right endpoint;
The screening module 840 is configured to screen the insurance claim material through a preset asynchronous thread to obtain a reference claim material with a material serial number and a target material class, where the material serial number is located in the target interval;
And the quality control module 850 is used for sequentially carrying out quality detection on the reference claim material through the target material class and the asynchronous thread according to the material serial number, and obtaining a quality control result of the quality detection until the material serial number is equal to the right end point of the interval, wherein the quality control result comprises that the quality detection is passed or the quality detection is failed.
The specific implementation of the quality control device based on the insurance claim is basically the same as the specific embodiment of the quality control method based on the insurance claim, and is not repeated here.
The embodiment of the application also provides electronic equipment, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the quality control method based on insurance claims when executing the computer program. The electronic equipment can be any intelligent terminal including a tablet personal computer, a vehicle-mounted computer and the like.
Referring to fig. 9, fig. 9 illustrates a hardware structure of an electronic device according to another embodiment, the electronic device includes:
The processor 910 may be implemented by a general purpose central processing unit (Central Processing Unit, CPU), a microprocessor, an Application SPECIFIC INTEGRATED Circuit (ASIC), or one or more integrated circuits, etc. for executing related programs to implement the technical solutions provided by the embodiments of the present application;
the Memory 920 may be implemented in the form of a Read Only Memory (ROM), a static storage device, a dynamic storage device, or a random access Memory (Random Access Memory, RAM). The memory 920 may store an operating system and other application programs, and when the technical solutions provided in the embodiments of the present disclosure are implemented by software or firmware, relevant program codes are stored in the memory 920, and the processor 910 invokes the quality control method based on insurance claims to execute the embodiments of the present disclosure;
an input/output interface 930 for inputting and outputting information;
The communication interface 940 is configured to implement communication interaction between the device and other devices, and may implement communication in a wired manner (e.g., USB, network cable, etc.), or may implement communication in a wireless manner (e.g., mobile network, WIFI, bluetooth, etc.);
A bus 950 for transferring information between components of the device (e.g., processor 910, memory 920, input/output interface 930, and communication interface 940);
Wherein processor 910, memory 920, input/output interface 930, and communication interface 940 implement communication connections among each other within the device via a bus 950.
The embodiment of the application also provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program which realizes the quality control method based on insurance claims when being executed by a processor.
The memory, as a non-transitory computer readable storage medium, may be used to store non-transitory software programs as well as non-transitory computer executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory remotely located relative to the processor, the remote memory being connectable to the processor through a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
According to the quality control method based on the insurance claim, the quality control device based on the insurance claim, the electronic equipment and the computer storage medium, the quality control process is free from manual intervention by carrying out automatic quality control on the insurance claim material, and the quality control efficiency of the insurance claim is improved.
The embodiments described in the embodiments of the present application are for more clearly describing the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application, and those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.
It will be appreciated by persons skilled in the art that the embodiments of the application are not limited by the illustrations, and that more or fewer steps than those shown may be included, or certain steps may be combined, or different steps may be included.
The above described apparatus embodiments are merely illustrative, wherein the units illustrated as separate components may or may not be physically separate, i.e. may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
Those of ordinary skill in the art will appreciate that all or some of the steps of the methods, systems, functional modules/units in the devices disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof.
The terms "first," "second," "third," "fourth," and the like in the description of the application and in the above figures, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the application described herein may be implemented in sequences other than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "and/or" is used to describe an association relationship of an associated object, and indicates that three relationships may exist, for example, "a and/or B" may indicate that only a exists, only B exists, and three cases of a and B exist simultaneously, where a and B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one of a, b or c may represent a, b, c, "a and b", "a and c", "b and c", or "a and b and c", wherein a, b, c may be single or plural.
In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other manners. For example, the above-described apparatus embodiments are merely illustrative, and for example, the above-described division of units is merely a logical function division, and there may be another division manner in actual implementation, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be an indirect coupling or communication connection via some interfaces, devices or units, which may be in electrical, mechanical or other form.
The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units, may be located in one place, or may be distributed over a plurality of network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
In addition, each functional unit in the embodiments of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit. The integrated units may be implemented in hardware or in software functional units.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application may be embodied in essence or a part contributing to the prior art or all or part of the technical solution in the form of a software product stored in a storage medium, including multiple instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method of the various embodiments of the present application. The storage medium includes various media capable of storing programs, such as a U disk, a removable hard disk, a Read-Onl y Memory (ROM), a random access memory (Random Access Memory, RAM), a magnetic disk, or an optical disk.
The preferred embodiments of the present application have been described above with reference to the accompanying drawings, and are not thereby limiting the scope of the claims of the embodiments of the present application. Any modifications, equivalent substitutions and improvements made by those skilled in the art without departing from the scope and spirit of the embodiments of the present application shall fall within the scope of the claims of the embodiments of the present application.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510962763.2A CN120852064A (en) | 2025-07-11 | 2025-07-11 | Quality control method and device based on insurance claim, electronic equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510962763.2A CN120852064A (en) | 2025-07-11 | 2025-07-11 | Quality control method and device based on insurance claim, electronic equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN120852064A true CN120852064A (en) | 2025-10-28 |
Family
ID=97425386
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510962763.2A Pending CN120852064A (en) | 2025-07-11 | 2025-07-11 | Quality control method and device based on insurance claim, electronic equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120852064A (en) |
-
2025
- 2025-07-11 CN CN202510962763.2A patent/CN120852064A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9552511B2 (en) | Identifying images using face recognition | |
| JP6397144B2 (en) | Business discovery from images | |
| GB2402535A (en) | Face recognition | |
| Qi et al. | A convolutional neural network face recognition method based on BILSTM and attention mechanism | |
| CN113704623A (en) | Data recommendation method, device, equipment and storage medium | |
| CN114139630A (en) | Gesture recognition method and device, storage medium and electronic equipment | |
| CN112214626A (en) | Image recognition method and device, readable storage medium and electronic equipment | |
| CN117763126A (en) | Knowledge retrieval methods, equipment, storage media and devices | |
| WO2024027347A9 (en) | Content recognition method and apparatus, device, storage medium, and computer program product | |
| CN113495987B (en) | Data search method, device, equipment and storage medium | |
| CN116704581B (en) | Face recognition method, device, equipment and storage medium | |
| CN117037172B (en) | Text recognition method, text recognition device, computer device, and storage medium | |
| CN112380369A (en) | Training method, device, equipment and storage medium of image retrieval model | |
| CN120852064A (en) | Quality control method and device based on insurance claim, electronic equipment and storage medium | |
| CN116521873B (en) | Comment detection method and comment detection device, electronic equipment and storage medium | |
| US20240184860A1 (en) | Methods and arrangements for providing impact imagery | |
| CN116580442A (en) | Micro-expression recognition method, device, equipment and medium based on separable convolution | |
| CN116597245B (en) | An image recognition model training method and system, and an image processing method and system. | |
| CN119477382B (en) | Lead following method, lead following device, electronic device and storage medium | |
| CN114913556B (en) | Image recognition method, device, system and storage medium | |
| CN116486835B (en) | Synthetic speech detection method and system, computer equipment, and storage medium | |
| CN117037219B (en) | Palmprint recognition method, device, computer equipment, storage medium and program product | |
| CN116501871B (en) | User portrait construction method, device, equipment and medium based on multimodal data | |
| CN113128296B (en) | Electronic handwriting signature fuzzy label recognition system | |
| CN120822093A (en) | Task prediction method and device based on multimodal data, electronic device, and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |