Regional medical co-body information integrated management system and method based on cloud platform

Abstract

The invention relates to the field of intelligent management, in particular to a regional medical cobody information integrated management system and method based on a cloud platform. The system comprises a platform deployment management module, an operation maintenance management module, a basic setting management module and a medical insurance comprehensive management module, through the above modules, efficient sharing and cooperation of medical information in a region can be realized through a unified technical interface, the quality of a medical service system is improved, and thus more convenient and high-quality medical services are provided for patients.

Description

Technical Field

[0001] This application relates to the field of intelligent management, specifically to a cloud-based integrated information management system and method for regional medical consortia. Background Technology

[0002] Regional medical consortia, as a highly innovative healthcare service model, refer to a healthcare service network formed by one or more medical institutions within a specific region. This model aims to achieve collaborative cooperation among medical institutions at all levels through resource sharing, thereby improving the overall efficiency and quality of healthcare services and providing the public with more continuous and efficient healthcare.

[0003] However, under the traditional regional medical consortium management model, the information systems of each medical institution are built independently, lacking unified standards and technical interfaces. This makes data interoperability and mutual recognition difficult, affecting patient referral efficiency and service continuity. Furthermore, system operation and maintenance are complex, with each institution operating independently and lacking unified standards and specifications, making system updates and upgrades difficult and unable to adapt to rapidly changing medical business needs.

[0004] Therefore, a cloud-based integrated information management system for regional medical consortia is desired. Summary of the Invention

[0005] This application was made in consideration of the above problems. One object of this application is to provide a cloud-based integrated information management system and method for regional medical consortia.

[0006] The embodiments of this application provide a cloud-based regional medical consortium information integration management system, which includes: a platform deployment management module, an operation and maintenance management module, an infrastructure management module, and a comprehensive medical insurance management module; The platform deployment management module includes a system configuration unit, an organization maintenance unit, and an organization department personnel management unit. The system configuration unit is used to maintain system resources and system modules; the organization maintenance unit is used to maintain organization information and system module permissions. The operation and maintenance management module includes a log query unit, an interface call query unit, a system user management unit, a system parameter maintenance unit, and a parameter business template maintenance unit. The log query unit is used for log querying, the interface call query unit is used for interface call querying, the system user management unit is used for managing system users, the system parameter maintenance unit is used for maintaining system parameters, and the parameter business template maintenance unit is used for maintaining parameter business templates. The basic settings management module includes a disease diagnosis maintenance unit, a billing item maintenance unit, a medical order item maintenance unit, a commodity filing management unit, a commodity dictionary review unit, a supplier maintenance unit, a drug manufacturer maintenance unit, a medication route maintenance unit, a surgical dictionary maintenance unit, a payment method maintenance unit, an income classification maintenance unit, a system dictionary maintenance unit, an examination site dictionary maintenance unit, an examination method dictionary maintenance unit, and a test and examination mutual recognition code comparison unit. The medical insurance comprehensive management module includes a medical insurance drug comparison maintenance unit and a medical insurance cost comparison maintenance unit. The medical insurance drug comparison maintenance unit is used for drug comparison and material comparison, and the medical insurance cost comparison maintenance unit is used for diagnosis and treatment comparison and material comparison.

[0007] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the log query unit includes: The log query input subunit is used to obtain log query input from the user. The log report extraction subunit is used to extract various logs from the log database; The log report semantic embedding subunit is used to perform semantic embedding encoding on each log in the log database to obtain a set of log report semantic embedding encoding vectors; A log query input semantic embedding subunit is used to perform semantic embedding encoding on the log query input to obtain a log query intent embedding encoding vector. The log query dynamic semantic retrieval subunit is used to perform adaptive semantic neighborhood pruning on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector to obtain the log query semantic response encoding vector. The log semantic search result generation subunit is used to obtain log semantic search results representing the sequence number tags of the log based on the log query semantic response encoding vector.

[0008] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the log report semantic embedding subunit is used to: perform semantic embedding encoding on each log in the log database using a semantic embedding encoder containing a BERT model to obtain a set of log report semantic embedding encoding vectors.

[0009] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the log query input semantic embedding subunit is used to: perform semantic embedding encoding on the log query input using the semantic embedding encoder containing the BERT model to obtain the log query intent embedding encoding vector.

[0010] For example, according to an embodiment of the cloud-based regional medical consortium information integration management system of this application, the log query dynamic semantic retrieval subunit includes: The robust semantic anchoring subunit is used to determine a robust semantic anchoring vector representing the core query direction based on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector. The dynamic semantic optimization candidate second-level subunit is used to construct a dynamic semantic optimization candidate set based on the robust semantic anchor vector by comprehensively considering the similarity between the log report semantic embedding encoding vector and the robust semantic anchor vector, as well as their diversity in the semantic neighborhood; and... The semantic difference gating fusion secondary subunit is used to selectively fuse the semantic difference information contained in the dynamic semantic optimization candidate set into the robust semantic anchoring vector through a gating fusion mechanism to generate the log query semantic response encoding vector.

[0011] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the robust semantic anchoring secondary subunit is used for: Calculate the semantic similarity between the log query intent embedding encoding vector and each of the log report semantic embedding encoding vectors; A predetermined number of log report semantic embedding encoding vectors with the highest semantic similarity to the log query intent embedding encoding vector are selected to form an anchor point candidate set; and, The robust semantic anchoring vector is obtained by weighting and averaging the vectors in the anchor candidate set.

[0012] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the dynamic semantic optimization candidate secondary subunit is used for: A score for diversity within a semantic neighborhood is determined by calculating the average semantic distance metric between a candidate log report semantic embedding encoding vector and other log report semantic embedding encoding vectors within its semantic neighborhood; and, A comprehensive score for each candidate log report semantic embedding encoding vector is determined by a weighted sum of the similarity scores between the candidate log report semantic embedding encoding vector and the robust semantic anchor vector, and the scores of the diversity scores.

[0013] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the semantic difference gating fusion secondary subunit is used to calculate a gating value for controlling the fusion ratio of the semantic difference information, and the gating value is dynamically calculated based on the robust semantic anchoring vector and the aggregated difference information generated by the dynamic semantic optimization candidate set.

[0014] For example, in the cloud-based regional medical consortium information integration management system according to an embodiment of this application, the log semantic search result generation subunit is used to: input the log query semantic response encoding vector into a classifier-based semantic search engine to obtain the log semantic search result, wherein the log semantic search result is a log sequence number tag.

[0015] The embodiments of this application also provide a cloud-based method for integrated management of regional medical consortium information, wherein the cloud-based method for integrated management of regional medical consortium information is applied to any of the aforementioned cloud-based integrated management systems for regional medical consortium information.

[0016] The cloud-based regional medical consortium information integration management system and method according to the embodiments of this application can realize efficient sharing and collaboration of medical information within the region through a unified technical interface, improve the quality of the medical service system, and thus provide patients with more convenient and high-quality medical services. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings of the embodiments of this application will be briefly described below. Obviously, the drawings described below only relate to some embodiments of this application, and are not intended to limit this application.

[0018] Figure 1 This application shows a schematic diagram of the structure of a cloud-based integrated information management system for regional medical consortia. Figure 2 This illustration shows a structural diagram of the operation and maintenance management module of the cloud-based regional medical consortium information integrated management system in this application embodiment; Figure 3 This application illustrates a schematic diagram of the log query unit of a cloud-based integrated information management system for regional medical consortia, as shown in this embodiment; and Figure 4 This paper illustrates the structural diagram of the log query dynamic semantic retrieval subunit of the regional medical consortium information integration management system based on a cloud platform in an embodiment of this application. Detailed Implementation

[0019] The terminology used in this specification is that which is currently widely used in the art in consideration of the functionality of this application; however, these terms may vary depending on the intent of a person skilled in the art, precedent, or new technology in the art. Furthermore, specific terms may be chosen, and in such cases, their detailed meanings will be described in the detailed description of this application. Therefore, the terminology used in this specification should not be construed as simple names, but rather based on the meaning of the terms and the overall description of this application.

[0020] While this application makes various references to certain modules of the systems according to embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0021] This application uses flowcharts to illustrate the operations performed by the system according to embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously, as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0022] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are also within the scope of protection of this application.

[0023] Based on this, this application proposes a cloud-based integrated information management system for regional medical consortia. This system enables efficient sharing and collaboration of medical information within the region through a unified technical interface, improving the quality of the medical service system and thus providing patients with more convenient and higher-quality medical services. Specifically, such as... Figure 1 and Figure 2As shown, the cloud-based regional medical consortium information integration management system 100 includes: a platform deployment management module 110, an operation and maintenance management module 120, a basic settings management module 130, and a comprehensive medical insurance management module 140; wherein, the platform deployment management module 110 includes a system configuration unit, an institution maintenance unit, and an institution department personnel management unit, the system configuration unit being used to maintain system resources and system modules; the institution maintenance unit being used to maintain institution information and system module permissions; wherein, the operation and maintenance management module 120 includes a log query unit 121, an interface call query unit 122, a system user management unit 123, a system parameter maintenance unit 124, and a parameter business template maintenance unit 125, the log query unit 121 being used for log query, the interface call query unit 122 being used for interface call query, the system user management unit 123, the system parameter maintenance unit 124, and the ... system parameter maintenance unit 124, the system parameter business template maintenance unit 125, the system parameter business template maintenance unit 125, the system parameter business template maintenance unit 126, the system parameter business template maintenance unit 127, the system parameter business template maintenance unit 128, the system parameter business template maintenance unit 129, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 120, the system parameter business template maintenance unit 1 23 is used to manage system users; the system parameter maintenance unit 124 is used to maintain system parameters; the parameter business template maintenance unit 125 is used to maintain parameter business templates. The basic settings management module 130 includes a disease diagnosis maintenance unit, a fee item maintenance unit, a medical order item maintenance unit, a commodity filing management unit, a commodity dictionary review unit, a supplier maintenance unit, a drug manufacturer maintenance unit, a medication route maintenance unit, a surgical dictionary maintenance unit, a payment method maintenance unit, an income classification maintenance unit, a system dictionary maintenance unit, an examination site dictionary maintenance unit, an examination method dictionary maintenance unit, and a laboratory examination mutual recognition code comparison unit. The medical insurance comprehensive management module includes a medical insurance drug comparison maintenance unit and a medical insurance expense comparison maintenance unit. The medical insurance drug comparison maintenance unit is used for drug comparison and material comparison, and the medical insurance expense comparison maintenance unit is used for diagnosis and treatment comparison and material comparison.

[0024] The cloud-based regional medical consortium information integration management system aims to integrate medical resources and improve the efficiency and quality of medical services.

[0025] Specifically, in one example, within the platform deployment management module 110, the system configuration unit is responsible for configuring infrastructure such as servers, networks, and storage to ensure the high performance and stability of the system. It also includes the functional configuration of different system modules to meet the needs of various medical institutions. The institution maintenance unit is used to register and update the basic information of each medical institution and set the access permissions of these institutions in the system to ensure information security and compliance. The institution department personnel management unit is responsible for managing and allocating information of various departments and their staff within the medical institution, including role definitions and permission settings, to ensure that only authorized personnel can access sensitive data.

[0026] In the operation and maintenance management module 120, the log query unit 121 is used to record system operation logs, providing audit tracing functions to help troubleshoot problems and monitor security; the interface call query unit 122 is used to track API usage, analyze performance bottlenecks, and provide data support for service optimization; the system user management unit 123 is used to create, delete, and modify user accounts, set password policies and account locking mechanisms to ensure system security; the system parameter maintenance unit 124 is used to adjust system operating parameters, such as cache size and connection pool settings, to optimize system performance; and the parameter business template maintenance unit 125 is used to design and manage business process templates to ensure the consistency and repeatability of business rules.

[0027] In the basic settings management module 130, the disease diagnosis maintenance unit is used to input and update disease codes and descriptions, facilitating doctors to accurately select and record patient conditions; the billing item maintenance unit is used to establish billing standards, ensuring all fees are transparent and compliant, and preventing overcharging; the medical order item maintenance unit is used to standardize medical order formats, guiding medical staff to correctly prescribe treatment plans; the product filing management unit is used to record detailed information on medicines and other medical supplies, facilitating inventory management and procurement decisions; the product dictionary review unit is used to check the accuracy of product information, ensuring it matches the actual product and avoiding errors; the supplier maintenance unit is used to store supplier information, helping hospitals evaluate and select reliable partners; and the drug manufacturer maintenance unit is used to collect and maintain pharmaceutical company data, providing information for drug supply. Source tracing provides support; the medication route maintenance unit determines the drug administration method to assist doctors in making the best treatment choices; the surgical dictionary maintenance unit compiles a list of surgical types to simplify the recording process of surgical information; the payment method maintenance unit sets payment options to improve the patient's payment experience; the income classification maintenance unit defines income categories to facilitate financial statistics and report generation; the system dictionary maintenance unit builds a unified data dictionary to promote data exchange between different systems; the examination site dictionary maintenance unit clarifies the anatomical location corresponding to various examinations to reduce the risk of misdiagnosis; the examination method dictionary maintenance unit lists available diagnostic techniques to assist doctors in developing treatment plans; and the laboratory test mutual recognition code comparison unit enables the sharing of test results between different hospitals to reduce duplicate examinations.

[0028] In the medical insurance integrated management module 140, the medical insurance drug comparison and maintenance unit is used to connect the hospital drug database with the medical insurance reimbursement catalog to ensure that patients can enjoy the medical insurance benefits they are entitled to; the medical insurance expense comparison and maintenance unit is used to verify the cost of medical service items according to medical insurance policies to ensure accurate settlement.

[0029] Each unit should enable human-computer interaction via a graphical user interface (GUI) or command-line tool (CLI), while utilizing a database for data persistence. To ensure system flexibility and scalability, a microservices architecture should be adopted, and deployment should be carried out using containerization technologies (such as Docker) and orchestration tools (such as Kubernetes). Furthermore, all units must adhere to stringent security standards, including data encryption, authentication, and access control, to protect patient privacy and medical information security.

[0030] Furthermore, it's understandable that log querying plays a crucial role in healthcare information systems. When system malfunctions or exhibits abnormal behavior, log queries can quickly locate relevant error messages and operation records, helping maintenance personnel rapidly pinpoint the root cause of the problem. However, traditional log queries primarily rely on keyword matching, which is susceptible to spelling errors, synonyms, or incomplete matches, failing to capture complex semantic relationships and leading to inaccurate query results. Moreover, when performing full-text searches on large-scale log data, traditional query methods often require a considerable amount of time to return results, especially with massive log volumes.

[0031] Specifically, in the log query unit, the technical concept of this application is to obtain the log query input input by the user, extract each log from the log database, and use AI-based data analysis and processing technology to perform semantic embedding encoding on each log and the log query input respectively. This allows for the automatic generation of log sequence number tags by guiding a dynamic semantic response representation based on the selection range ratio between the log query intent embedding features and the semantic embedding features of each log report. Compared to traditional keyword matching, this application can understand and capture the deep semantic information behind the user query, accurately understand the intent, and quickly find the most matching log entry, thereby shortening the query time and improving response speed.

[0032] Accordingly, such as Figure 3As shown, the log query unit 121 includes: a log query input subunit 1211, used to obtain log query input input by the user; a log report extraction subunit 1212, used to extract various logs from the log database; a log report semantic embedding subunit 1213, used to perform semantic embedding encoding on various logs in the log database to obtain a set of log report semantic embedding encoding vectors; a log query input semantic embedding subunit 1214, used to perform semantic embedding encoding on the log query input to obtain a log query intent embedding encoding vector; a log query dynamic semantic retrieval subunit 1215, used to perform adaptive semantic neighborhood pruning feature dynamic semantic retrieval on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector to obtain a log query semantic response encoding vector; and a log semantic search result generation subunit 1216, used to obtain log semantic search results representing the sequence number tags of the logs based on the log query semantic response encoding vector. The log report semantic embedding subunit 1213 is configured to: perform semantic embedding encoding on each log in the log database using a semantic embedding encoder containing a BERT model to obtain a set of log report semantic embedding encoding vectors. The log query input semantic embedding subunit 1214 is configured to: perform semantic embedding encoding on the log query input using the semantic embedding encoder containing a BERT model to obtain the log query intent embedding encoding vector.

[0033] Specifically, in the technical solution of this application, firstly, the log query input provided by the user is obtained. Then, various logs are extracted from the log database.

[0034] Then, considering that in actual log query scenarios, users often do not simply search for logs based on keywords, but rather expect to retrieve content that matches their intent from a semantic perspective. For example, a user wants to query "system operation records related to abnormal medical insurance reimbursement." Here, "abnormal medical insurance reimbursement" involves a deeper semantic understanding, and it is difficult to accurately locate relevant logs by simply relying on word matching on the surface of the text. Based on this, this application performs semantic embedding encoding on each log in the log database to transform the log information into a vector form containing rich semantic information, thereby obtaining a set of log report semantic embedding encoding vectors. In particular, in a specific embodiment of this application, a semantic embedding encoder containing a BERT model is used to perform semantic embedding encoding on each log in the log database to use the BERT model to capture the semantic relationships between these words and sentences, understand their deeper meaning, and thus mine more accurate and comprehensive semantic features of the log content, thereby obtaining the set of log report semantic embedding encoding vectors.

[0035] Similarly, considering that in log query scenarios, users often input their query intent in a natural language description, the content of which contains rich semantic information. For example, if a user inputs "query logs of abnormal interface calls in yesterday's operation and maintenance management module," it is difficult to accurately locate the relevant logs simply by matching the text surface. Based on this, in the technical solution of this application, the log query input is semantically embedded and encoded to mine deeper semantics, resulting in a log query intent embedding encoding vector. In particular, in a specific embodiment of this application, the semantic embedding encoder containing the BERT model is used to semantically embed and encode the log query input to mine these implicit semantic relationships, which are then integrated into the query intent embedding encoding vector to obtain the log query intent embedding encoding vector. This provides rich data representation for subsequent semantic queries.

[0036] Furthermore, considering that in log query scenarios, both the user's query intent and the log content itself often possess complex semantic structures. For example, a user might want to query "logs related to abnormal maintenance of billing items in the basic settings management module within the past week, but which are also related to medical insurance reimbursement." This involves multiple semantic information aspects, including time range, specific module operations, abnormal situations, and connections with other businesses. Simply relying on traditional semantic similarity calculation methods (such as simple cosine similarity calculation) may lead to misjudgments or inaccurate matching when faced with massive and complex log data and diverse query intents. Therefore, this application proposes an adaptive semantic neighborhood pruning feature-dynamic semantic retrieval method to perform query responses on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector to obtain the log query semantic response encoding vector. In other words, when performing semantic retrieval in a massive, high-dimensional log vector space, it is necessary to overcome the contradiction between the ambiguity of user query intent and the diversity of log content. Specifically, it is difficult to ensure high relevance while avoiding semantic homogenization of search results, and it is also difficult to dynamically adjust the degree of adoption of candidate log information. As a result, the final generated response vector cannot accurately and comprehensively express the complex relationship between query intent and log context. Accordingly, such as Figure 4As shown, the log query dynamic semantic retrieval subunit 1215 includes: a robust semantic anchoring subunit 12151, used to determine a robust semantic anchoring vector representing the core query direction based on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector; a dynamic semantic optimization candidate subunit 12152, used to construct a dynamic semantic optimization candidate set based on the robust semantic anchoring vector by comprehensively considering the similarity between the log report semantic embedding encoding vector and the robust semantic anchoring vector and their diversity in the semantic neighborhood; and a semantic difference gating fusion subunit 12153, used to selectively fuse the semantic difference information contained in the dynamic semantic optimization candidate set into the robust semantic anchoring vector through a gating fusion mechanism to generate the log query semantic response encoding vector.

[0037] Specifically, the robust semantic anchoring secondary subunit 12151 embeds an encoding vector based on the log query intent, for example, denoted as... The set of semantic embedding encoding vectors of the log report, for example denoted as ,in It is a log index used to determine a robust semantic anchor vector representing the core query direction, for example, denoted as Here, the received log query intent is embedded with an encoded vector. With each of the log report semantic embedding encoding vectors First, the semantic similarity between them is calculated to obtain a set of similarity scores. For example, cosine similarity calculation can be used. Then, from the set of semantic embedding encoded vectors in the log report... Selected from The top-K log report semantic embedding encoding vectors with the highest similarity scores are used to form the initial anchor candidate set. ,(in ), and their corresponding scores The K scores are then subjected to Softmax normalization to obtain the weight of each candidate vector. Finally, by calculating the weighted average of these K candidate vectors, a more robust semantic anchor vector that better represents the center of the core semantic region is generated than a single highest-scoring vector. As a benchmark for subsequent dynamic search, it has better robustness than choosing a single most similar vector as an anchor point.

[0038] That is, the robust semantic anchoring secondary subunit 12151 is used to: calculate the semantic similarity between the log query intent embedding encoding vector and each of the log report semantic embedding encoding vectors; select a preset number of log report semantic embedding encoding vectors with the highest semantic similarity to the log query intent embedding encoding vector to form an anchor candidate set; and perform a weighted average on each vector in the anchor candidate set to obtain the robust semantic anchoring vector.

[0039] Furthermore, the dynamic semantic optimization candidate second-level subunit 12152 performs an adaptive semantic neighborhood pruning mechanism, which is based on robust semantic anchoring vectors. and the set of semantic embedding encoding vectors of the log report Determine the candidate set for dynamic semantic optimization, for example, denoted as First, for the input robust semantic anchor vector... and the set of semantic embedding encoding vectors of the log report Perform preliminary screening and calculate robust semantic anchoring vectors. The similarity, such as cosine similarity, between each log report semantic embedding encoding vector is used to filter out all similarities greater than a preset threshold, for example, denoted as . The vectors form the initial neighborhood. Then, perform diversity pruning, that is, for Each vector in the vector, for example denoted as Calculate its diversity score within the neighborhood. Here, the diversity score is defined as the average cosine distance between the vector and all other vectors in its neighborhood, which is one minus the cosine similarity: ;in For the initial neighborhood The number of vectors within, and The second norm of a vector.

[0040] Therefore, All vectors in can be determined based on their relationship with Similarity score and its own diversity score To perform a comprehensive ranking, for example, you can define a comprehensive score: ;in and The weights are adjustable. Furthermore, based on the overall score, the top M vectors are selected from highest to lowest, or vectors with scores exceeding a certain dynamic threshold are selected to form the final dynamic semantic optimization candidate set. It is not only highly relevant to the query anchor, but also possesses semantic diversity.

[0041] That is, the dynamic semantic optimization candidate second-level subunit 12152 is used to: determine the diversity score in the semantic neighborhood by calculating the average semantic distance metric between the candidate log report semantic embedding encoding vector and other log report semantic embedding encoding vectors in its semantic neighborhood; and determine the comprehensive score of each candidate log report semantic embedding encoding vector based on the weighted sum of the similarity between the candidate log report semantic embedding encoding vector and the robust semantic anchor vector and the diversity score.

[0042] Furthermore, the semantic difference gating fusion secondary subunit 12153 executes a gating fusion mechanism, which receives the log query intent embedding encoding vector. Robust semantic anchor vectors and dynamic semantic optimization candidate set Each of the dynamic semantic optimization candidate vectors is denoted as . And firstly, calculate each dynamic semantic optimization candidate vector. Relative to query vector semantic difference vector Then, each dynamic semantic optimization candidate vector is computed. Attention weights are assigned to reflect their importance to the query: ;in It is a learnable weight matrix, which makes attention calculation more flexible.

[0043] Then, based on the attention weights, all semantic difference vectors are weighted and summed to obtain the aggregated difference vector, for example, denoted as... .

[0044] Then the gating mechanism is executed, that is, the gating value is calculated. It is a scalar between 0 and 1, used to dynamically control the aggregation difference vector. The degree of fusion. This gating value is determined by the anchor vector. and aggregated difference vector Joint decision: ;in It's the Sigmoid function; [;] indicates vector concatenation. and These are the learnable parameters of the gating network.

[0045] Finally, aggregate the difference vectors. Through the gate value Integrate In the process, the final log query semantic response encoding vector is generated. : Thus, the log query semantic response encoding vector It intelligently absorbs key difference information from candidate logs while retaining a robust semantic core.

[0046] That is, the semantic difference gating fusion secondary subunit 12153 is used to calculate a gating value for controlling the fusion ratio of the semantic difference information, and the gating value is dynamically calculated based on the robust semantic anchoring vector and the aggregated difference information generated by the dynamic semantic optimization candidate set.

[0047] In summary, compared to traditional semantic retrieval methods that typically use a fixed number (Top-K) or fixed radius neighborhood for searching, the log query dynamic semantic retrieval subunit 1215, through an adaptive semantic neighborhood pruning mechanism, not only considers the similarity to the query intent but also introduces a diversity measure. It can dynamically construct a candidate set that is both highly relevant and semantically rich based on the density distribution of log semantics in the vector space, avoiding information redundancy caused by excessive log homogeneity or the omission of key weakly relevant semantic information due to inappropriate retrieval scope. This solves the problem of poor adaptability of traditional fixed-range retrieval methods under complex and fuzzy queries. Furthermore, a gated semantic fusion mechanism is used to weight and fuse the difference vectors. A learnable gating unit dynamically calculates the fusion coefficients to intelligently control the fusion ratio of differentiated information and anchor information. That is, this gating mechanism is similar to a gated recurrent unit (GRU), which can determine the value of the differentiated information provided by the current candidate set in optimizing the query response, thereby achieving precise control of the final response vector and preventing semantic drift of the response vector due to the introduction of noise or irrelevant semantic differences.

[0048] Therefore, through adaptive neighborhood pruning, a more representative set of candidate logs can be matched for fuzzy or complex queries, significantly improving the robustness and accuracy of retrieval. Furthermore, through the gating fusion mechanism, the most critical semantic differences can be intelligently incorporated into the response vector while filtering out noise information, so that the final response vector retains the core semantics and is rich in context, making it more expressive. Thus, based on the dynamic adaptive characteristics of anchor point positioning, neighborhood selection and final fusion, it can better adapt to different query intentions and log data distributions, realizing dynamic adaptive retrieval.

[0049] Furthermore, the log semantic search result generation subunit 1216 is used to: input the log query semantic response encoding vector into a classifier-based semantic search engine to obtain the log semantic search result, wherein the log semantic search result is the sequence number tag of the log.

[0050] It should be understood that the log query semantic response encoding vector is input into a classifier-based semantic search engine to obtain log semantic search results, where the log semantic search results are the log's sequence number tags. That is, the log query semantic response encoding vector obtained through dynamic semantic search using the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector is then classified. This leverages the classifier's powerful data feature discrimination and classification capabilities, accurately classifying the logs into different categories based on the features contained in the vector and the classification rules learned during training, thereby automatically obtaining the log's sequence number tag. In this way, operations and maintenance personnel can quickly and accurately locate specific log records from the log database based on this sequence number tag, enabling them to take appropriate measures to optimize the system, improve business processes, and ensure the stable and accurate operation of related business functions in the future.

[0051] Specifically, the log semantic search result generation subunit 1216 inputs the log query semantic response encoding vector obtained through dynamic semantic retrieval into a classifier-based semantic search engine to obtain specific log semantic search results. These results are presented in the form of log sequence number tags. A sequence number tag is essentially one or a set of unique identifiers associated with a specific log record; these can be used to quickly locate and access specific log entries in the database. For example, in the information system of a large medical institution, a large number of operational logs may be generated daily, recording various activities such as updates to patient information, diagnostic processes, and the execution of treatment plans. Each log entry has its unique sequence number tag, which is automatically generated by the system and ensures its uniqueness throughout the entire log database. When operations and maintenance personnel need to find a specific log record, they can input relevant query conditions (such as date range, event type, involved department or doctor, etc.) through the log semantic search function. The system will perform semantic analysis based on these conditions to find the most matching log query semantic response encoding vector.

[0052] Furthermore, this application also provides a cloud-based integrated management method for regional medical consortium information, wherein the cloud-based integrated management method for regional medical consortium information is applied to the cloud-based integrated management system for regional medical consortium information as described in the preceding claim.

[0053] Those skilled in the art will understand that all or part of the steps in the above methods can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiments can be implemented in hardware or as a software functional module. This application is not limited to any particular combination of hardware and software.

[0054] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. It should also be understood that terms such as those defined in a common dictionary shall be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or highly formalized meaning, unless expressly defined herein.

[0055] The above is a description of this application and should not be considered as a limitation thereof. Although several exemplary embodiments of this application have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of this application.

Claims

1. A cloud-based integrated information management system for regional medical consortia, characterized in that, include: The platform includes a deployment management module, an operation and maintenance management module, a basic settings management module, and a comprehensive medical insurance management module. The platform deployment management module includes a system configuration unit, an organization maintenance unit, and an organization department personnel management unit. The system configuration unit is used to maintain system resources and system modules; the organization maintenance unit is used to maintain organization information and system module permissions. The operation and maintenance management module includes a log query unit, an interface call query unit, a system user management unit, a system parameter maintenance unit, and a parameter business template maintenance unit. The log query unit is used for log querying, the interface call query unit is used for interface call querying, the system user management unit is used for managing system users, the system parameter maintenance unit is used for maintaining system parameters, and the parameter business template maintenance unit is used for maintaining parameter business templates. The basic settings management module includes a disease diagnosis maintenance unit, a billing item maintenance unit, a medical order item maintenance unit, a commodity filing management unit, a commodity dictionary review unit, a supplier maintenance unit, a drug manufacturer maintenance unit, a medication route maintenance unit, a surgical dictionary maintenance unit, a payment method maintenance unit, an income classification maintenance unit, a system dictionary maintenance unit, an examination site dictionary maintenance unit, an examination method dictionary maintenance unit, and a test and examination mutual recognition code comparison unit. The medical insurance comprehensive management module includes a medical insurance drug comparison maintenance unit and a medical insurance cost comparison maintenance unit. The medical insurance drug comparison maintenance unit is used for drug comparison and material comparison, and the medical insurance cost comparison maintenance unit is used for diagnosis and treatment comparison and material comparison.

2. The integrated information management system for regional medical consortia based on a cloud platform according to claim 1, characterized in that, The log query unit includes: The log query input subunit is used to obtain log query input from the user. The log report extraction subunit is used to extract various logs from the log database; The log report semantic embedding subunit is used to perform semantic embedding encoding on each log in the log database to obtain a set of log report semantic embedding encoding vectors; A log query input semantic embedding subunit is used to perform semantic embedding encoding on the log query input to obtain a log query intent embedding encoding vector. The log query dynamic semantic retrieval subunit is used to perform adaptive semantic neighborhood pruning on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector to obtain the log query semantic response encoding vector. The log semantic search result generation subunit is used to obtain log semantic search results representing the sequence number tags of the log based on the log query semantic response encoding vector.

3. The integrated information management system for regional medical consortia based on a cloud platform according to claim 2, characterized in that, The log report semantic embedding subunit is used to: use a semantic embedding encoder containing a BERT model to perform semantic embedding encoding on each log in the log database to obtain a set of log report semantic embedding encoding vectors.

4. The integrated information management system for regional medical consortia based on a cloud platform according to claim 3, characterized in that, The log query input semantic embedding subunit is used to: perform semantic embedding encoding on the log query input using the semantic embedding encoder containing the BERT model to obtain the log query intent embedding encoding vector.

5. The integrated information management system for regional medical consortia based on a cloud platform according to claim 4, characterized in that, The log query dynamic semantic retrieval subunit includes: The robust semantic anchoring subunit is used to determine a robust semantic anchoring vector representing the core query direction based on the set of the log query intent embedding encoding vector and the log report semantic embedding encoding vector. The dynamic semantic optimization candidate second-level subunit is used to construct a dynamic semantic optimization candidate set based on the robust semantic anchor vector by comprehensively considering the similarity between the log report semantic embedding encoding vector and the robust semantic anchor vector, as well as their diversity in the semantic neighborhood; and... The semantic difference gating fusion secondary subunit is used to selectively fuse the semantic difference information contained in the dynamic semantic optimization candidate set into the robust semantic anchoring vector through a gating fusion mechanism to generate the log query semantic response encoding vector.

6. The regional medical consortium information integration management system based on a cloud platform according to claim 5, characterized in that, The robust semantic anchoring secondary subunit is used for: Calculate the semantic similarity between the log query intent embedding encoding vector and each of the log report semantic embedding encoding vectors; A preset number of log report semantic embedding encoding vectors with the highest semantic similarity to the log query intent embedding encoding vector are selected to form an anchor point candidate set; as well as, The robust semantic anchoring vector is obtained by weighting and averaging the vectors in the anchor candidate set.

7. The regional medical consortium information integration management system based on a cloud platform according to claim 5, characterized in that, The dynamic semantic optimization candidate second-level subunit is used for: A score for diversity within a semantic neighborhood is determined by calculating the average semantic distance metric between a candidate log report semantic embedding encoding vector and other log report semantic embedding encoding vectors within its semantic neighborhood; and, A comprehensive score for each candidate log report semantic embedding encoding vector is determined by a weighted sum of the similarity scores between the candidate log report semantic embedding encoding vector and the robust semantic anchor vector, and the scores of the diversity scores.

8. The integrated information management system for regional medical consortia based on a cloud platform according to claim 5, characterized in that, The semantic difference gating fusion secondary subunit is used to calculate a gating value for controlling the fusion ratio of the semantic difference information, and the gating value is dynamically calculated based on the robust semantic anchoring vector and the aggregated difference information generated by the dynamic semantic optimization candidate set.

9. The integrated information management system for regional medical consortia based on a cloud platform according to claim 8, characterized in that, The log semantic search result generation subunit is used to: input the log query semantic response encoding vector into a classifier-based semantic search engine to obtain the log semantic search result, wherein the log semantic search result is the log sequence number tag.

10. A method for integrated information management of regional medical consortia based on a cloud platform, characterized in that, The cloud-based regional medical consortium information integration management method is applied to the cloud-based regional medical consortium information integration management system as described in any one of claims 1 to 9.

Images