Search

CN-122025057-A - Internet-based medical insurance service platform comprehensive intelligent management system

CN122025057ACN 122025057 ACN122025057 ACN 122025057ACN-122025057-A

Abstract

The invention discloses an Internet-based medical insurance service platform comprehensive intelligent management system, which relates to the technical field of medical insurance intelligent management and comprises a data receiving module, a portrait construction module, a strategy adaptation module, a compliance verification module and a service navigation module. The system performs heterogeneous feature cooperative processing on a user identity authentication sequence and a medical service application text to generate a user service requirement image, performs rule matrix matching through a multi-stage policy matching engine, outputs a hierarchical policy matching result containing matching degree distribution, performs bidirectional compliance verification on a cost settlement bill image and a medicine catalog query log based on the distribution to form a medical insurance settlement compliance report, and generates and pushes a personalized service navigation policy. The invention realizes deep integration depiction of the user service requirements and dynamic accurate compliance examination based on quantitative evaluation.

Inventors

  • TANG WEI
  • LU DASHUAI

Assignees

  • 山东微医互联网大健康集团有限公司

Dates

Publication Date
20260512
Application Date
20260205

Claims (10)

  1. 1. Internet-based medical insurance service platform comprehensive intelligent management system is characterized by comprising: the data receiving module is used for receiving full-volume medical insurance service interaction data from the heterogeneous medical insurance service terminals, wherein the full-volume medical insurance service interaction data comprises a user identity authentication sequence, a medical service application text, a fee settlement receipt image and a medicine catalog inquiry log; the portrait construction module performs cross-domain identity verification processing on the user identity authentication sequence to generate an identity verification feature vector, performs intention deep analysis processing on the medical service application text to generate a medical intention analysis map, and performs heterogeneous feature cooperative processing on the identity verification feature vector and the medical intention analysis map to generate a user service requirement portrait; the policy adaptation module is used for calling a multi-stage medical insurance policy adaptation engine to perform rule matrix matching processing on the user service requirement portrait so as to generate a hierarchical policy adaptation result set, wherein the hierarchical policy adaptation result set comprises matching degree distribution between different medical insurance policy clauses and current medical requirements; The compliance verification module is used for carrying out bidirectional compliance verification processing on the expense settlement receipt image and the medicine catalog inquiry log based on the matching degree distribution so as to generate a medical insurance settlement compliance report; And the service navigation module generates a personalized service navigation strategy according to the medical insurance settlement compliance report, and pushes the personalized service navigation strategy to a user interaction interface to guide a service flow.
  2. 2. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 1, wherein the cross-domain identity verification processing of the user identity authentication sequence to generate an identity verification feature vector includes: performing multi-source authentication data decomposition processing on the user identity authentication sequence, and extracting independent authentication characteristics from a public security household registration library, a social security card chip and biological recognition equipment; invoking a distributed authentication fusion model to perform space-time consistency check processing on the independent authentication features to generate a preliminary fusion authentication vector; The authentication history behavior pattern of the current user is obtained from a national medical insurance authentication center, the authentication history behavior pattern is input into an anti-fake detection neural network to perform abnormal pattern recognition, and the recognized normal pattern features are subjected to feature binding with the preliminary fusion authentication vector to generate the enhanced authentication feature vector; And carrying out weight distribution fusion processing on the enhanced authentication feature vectors of different authentication channels to generate the identity verification feature vector.
  3. 3. The internet-based medical insurance service platform integrated intelligent management system according to claim 1, wherein the performing the intention deep parsing process on the medical service application text to generate a medical intention parsing map includes: medical entity extraction processing is carried out on the medical service application text, and a disease description entity, a treatment scheme entity and a medical institution entity are identified; Invoking a medical knowledge graph embedding model to perform relationship reasoning processing on the disease description entity, the treatment scheme entity and the medical institution entity, and constructing an initial medical relationship network; Calculating the most-phrase meaning path from each entity node to a preset medical target node in the initial medical relation network, screening out a key intention transmission link based on path weight, and carrying out structural encapsulation on the key intention transmission link to form an intention path set; And mapping and aligning the intention path set with a medical standard flow library to generate the medical intention analysis map containing multiple layers of intention nodes.
  4. 4. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 1, wherein the heterogeneous feature co-processing of the identity verification feature vector and the medical intent resolution map to generate a user service requirement representation comprises: performing map structure adaptation processing on the identity verification feature vector, and converting the identity verification feature vector into node feature representation compatible with the medical intention analysis map; Establishing an identity-intention correlation attention mechanism, and calculating semantic correlation strength between each dimension feature in the identity verification feature vector and each intention node in the medical intention analysis map; performing personalized weighting processing on the medical intention analysis map based on the semantic association strength to generate a personalized intention map; Carrying out multi-mode feature fusion processing on the node features in the personalized intention map and the converted node feature representations to generate a fusion user feature matrix; and performing feature dimension reduction and structural recombination processing on the fused user feature matrix to form the user service requirement portrait comprising identity attribute, medical intention and service preference.
  5. 5. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 1, wherein the invoking the multi-level medical insurance policy adaptation engine to perform rule matrix matching processing on the user service requirement representation to generate a set of level policy adaptation results comprises: Inputting the user service requirement portrait into a policy dimension decomposition layer of the multi-stage medical insurance policy adaptation engine, and separating a participation type dimension, a medical treatment area dimension and a disease type dimension; for each separated dimension, invoking a corresponding policy rule sub-engine to perform multi-level rule matching processing to generate an initial policy matching set; the priority ordering and conflict resolution processing are carried out on the initial policy matching set through a rule conflict detection layer of the multistage medical insurance policy adaptation engine, and a conflict-free policy matching sequence is generated; And carrying out applicability scoring processing on each policy rule in the conflict-free policy matching sequence by using a policy weight computing model to generate the grading policy adaptation result set, wherein each adaptation result comprises a policy clause number, a matching degree score and an application range description.
  6. 6. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 1, wherein the performing bidirectional compliance verification processing on the expense settlement document image and the medicine catalog query log based on the matching degree distribution to generate a medical insurance settlement compliance report comprises: Determining a priority sequence of the expense items to be checked and a medicine use checking key point according to the matching degree distribution; carrying out structured information extraction processing on the expense settlement receipt image to obtain expense detail list data, and carrying out medication behavior analysis processing on the medicine catalog inquiry log to obtain medicine use mode data; calling a medical insurance compliance verification model to perform cross-validation processing on the expense list data and the medicine use mode data, and identifying a suspected illegal item set; Performing manual audit rule auxiliary judgment processing on the suspected illegal item set, wherein the manual audit rule auxiliary judgment processing comprises the steps of performing similarity matching on the suspected illegal item and a historical audit case library, acquiring a final audit conclusion of a matching case as a reference basis, and performing comprehensive judgment by combining with special rules of a current medical insurance policy; And integrating the verification results and the judgment basis of all the projects, and generating a medical insurance settlement compliance report comprising a compliance project list, a violation project list and judgment reasons, wherein the medical insurance settlement compliance report is used for identifying reimbursement qualification of the expense projects and compliance state of medicine use.
  7. 7. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 6, wherein the processing of extracting structured information from the fee settlement document image to obtain fee detail list data includes: performing multi-angle perspective correction processing on the expense settlement bill image to obtain a standardized bill image; carrying out full-field recognition processing on the standardized document image by a mixed text recognition engine, and extracting original text information; Invoking a medical fee semantic analyzer to conduct item classification processing on the original text information, and classifying fee items into at least one standard category of examination fee, treatment fee, medicine fee, material fee, operation fee and nursing fee; carrying out numerical value standardization treatment on the classified expense items, and unifying an amount unit and a calculation mode; And mapping the structured expense information with a medical insurance item coding library to generate the machine-readable expense detail list data.
  8. 8. The internet-based medical insurance service platform integrated management system according to claim 6, wherein the performing a medication behavior analysis process on the medicine catalog query log to obtain medicine usage pattern data includes: analyzing a time stamp sequence, a medicine coding sequence and a user operation type in the medicine catalog inquiry log; constructing a time-medicine two-dimensional analysis matrix, and identifying a medicine inquiry frequency mode and an associated inquiry mode of a user; Invoking a medication rationality evaluation model to analyze the medicine inquiry frequency mode and judging whether abnormal high-frequency inquiry or unconventional medication combination inquiry exists or not; the correlation between the query medicine and the illness state is evaluated by combining the historical diagnosis and treatment record of the user and the current illness diagnosis; and integrating the query mode analysis result, the rationality evaluation result and the relevance evaluation result to form the medicine use mode data.
  9. 9. The internet-based medical insurance service platform comprehensive intelligent management system according to claim 1, wherein the generating a personalized service navigation policy according to the medical insurance settlement compliance report includes: Analyzing a compliance item list and a violation item list in the medical insurance settlement compliance report, and determining the current use situation of medical insurance services of a user; Matching a service navigation template matched with the user service demand portrait and the current situation of medical insurance service use from a service strategy knowledge base; Based on the real-time geographic position of the user and the state data of the service mechanism, carrying out dynamic path optimization processing on the service navigation template to generate an initial navigation strategy; The user feedback learning mechanism is called to carry out personalized adjustment processing on the initial navigation strategy, wherein the personalized adjustment processing comprises the steps of analyzing historical service selection preference of a user, predicting the acceptance degree of the user for guiding different types of services, and adjusting the presentation mode and the recommendation sequence of the navigation strategy according to a prediction result; the personalized service navigation policy is generated that contains next service suggestions, potential risk cues, and alternative recommendations.
  10. 10. The system of claim 4, wherein the establishing an identity-intent associated attention mechanism calculates a semantic association strength between each dimension feature in the identity verification feature vector and each intent node in the medical intent resolution graph, comprising: inputting the identity verification feature vector and the feature vectors of all intention nodes in the medical intention analysis map into a multi-head attention network; performing linear transformation on the identity verification feature vector to generate a query vector, and performing linear transformation on the feature vector of the intention node to generate a key vector and a value vector; Performing dot product calculation on the query vector and each key vector to obtain an initial attention score; scaling the initial attention score, normalizing by using a Softmax function, and generating the semantic association strength; And carrying out weighted summation on the value vectors corresponding to the intention nodes according to the semantic association strength to generate the context-enhanced identity-related intention representation.

Description

Internet-based medical insurance service platform comprehensive intelligent management system Technical Field The invention belongs to the technical field of medical insurance intelligent management, and particularly relates to an Internet-based medical insurance service platform comprehensive intelligent management system. Background The mapping from heterogeneous data sources to unified semantic expression is realized by carrying out collaborative fusion processing on the structural feature vector generated by cross-domain identity verification and the medical intention map generated by text deep analysis. This process enables extraction of potential association patterns between identity attribute fields and medical intent fields, such that the resulting demand portraits contain not only ground truth labels, but also deep features derived from the association patterns that are predictive. The technical scheme enables the understanding of the system to the user's needs to be changed from discrete fact lists to coherent and context-explanatory overall descriptions. By designing the output of the multi-level policy adaptation engine to include a distribution set of term matching metric values, a continuous, differentiable decision basis is provided for subsequent flows. Based on the quantitative distribution, the system has state sensing capability for verification logic of settlement document images and medicine query logs. The system can dynamically allocate computing resources according to matching strength or conflict conditions of different policy dimensions in the distribution, and adjust association closeness of verification depths and verification logics of different data sources. The technical scheme changes the compliance verification process from judgment based on a fixed threshold value to a continuous reasoning process which is based on probability evaluation and can be adaptively adjusted. Disclosure of Invention The present invention aims to solve at least one of the technical problems existing in the prior art; Therefore, the invention provides an Internet-based medical insurance service platform comprehensive intelligent management system, which comprises: the data receiving module is used for receiving full-volume medical insurance service interaction data from the heterogeneous medical insurance service terminals, wherein the full-volume medical insurance service interaction data comprises a user identity authentication sequence, a medical service application text, a fee settlement receipt image and a medicine catalog inquiry log; the portrait construction module performs cross-domain identity verification processing on the user identity authentication sequence to generate an identity verification feature vector, performs intention deep analysis processing on the medical service application text to generate a medical intention analysis map, and performs heterogeneous feature cooperative processing on the identity verification feature vector and the medical intention analysis map to generate a user service requirement portrait; the policy adaptation module is used for calling a multi-stage medical insurance policy adaptation engine to perform rule matrix matching processing on the user service requirement portrait so as to generate a hierarchical policy adaptation result set, wherein the hierarchical policy adaptation result set comprises matching degree distribution between different medical insurance policy clauses and current medical requirements; The compliance verification module is used for carrying out bidirectional compliance verification processing on the expense settlement receipt image and the medicine catalog inquiry log based on the matching degree distribution so as to generate a medical insurance settlement compliance report; And the service navigation module generates a personalized service navigation strategy according to the medical insurance settlement compliance report, and pushes the personalized service navigation strategy to a user interaction interface to guide a service flow. Further, the cross-domain identity verification processing of the user identity authentication sequence to generate an identity verification feature vector includes: performing multi-source authentication data decomposition processing on the user identity authentication sequence, and extracting independent authentication characteristics from a public security household registration library, a social security card chip and biological recognition equipment; invoking a distributed authentication fusion model to perform space-time consistency check processing on the independent authentication features to generate a preliminary fusion authentication vector; The authentication history behavior pattern of the current user is obtained from a national medical insurance authentication center, the authentication history behavior pattern is input into an anti-fake detection neural network to perform abnorm