Medical big data sharing platform based on cloud computing

Abstract

The invention discloses a medicine big data sharing platform based on cloud computing, and belongs to the field of data management, and the platform comprises a network construction unit which is used for building a distributed storage network on a cloud computing center; the interface setting unit is used for setting a mechanism access interface and a personal access interface; the database forming unit is used for forming a medicine big database; the mechanism request processing unit is used for screening an initial medicine data set according to the mechanism access request; the mechanism data processing unit is used for generating a target medicine data set and sending the target medicine data set to the requesting mechanism; the personal request processing unit is used for screening an associated medicine data set according to the personal access request; and the individual report generation unit is used for generating a personalized medicine data analysis report and sending the report to the requesting individual. By integrating and storing dispersed medical data and providing data access interfaces and processing services oriented to organizations and individuals, efficient sharing and utilization of the medical data are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data management, in particular to a cloud computing-based medical big data sharing platform. BACKGROUND

[0002] With the continuous development of medical informatization, a large amount of medical data is generated and stored by various medical institutions, including electronic medical records, medical images, test reports and other multi-dimensional data resources. However, at present, these medical data are mostly scattered and stored in the local databases of various medical institutions, lacking a unified sharing platform and mechanism, resulting in difficulty in efficient sharing and utilization of medical data among medical institutions. In addition, in the face of the complexity and diversity of medical big data, different data demand subjects, such as medical institutions, research institutions, pharmaceutical enterprises and individual users, often need to analyze data in different ways. The existing technology lacks flexible data processing mechanism and cannot provide personalized data analysis services according to the needs of different subjects, making it difficult to meet the diversified application needs of medical data. SUMMARY

[0003] The present application provides a cloud computing-based medical big data sharing platform, aiming to solve the technical problem that medical data are scattered and stored in various medical institutions, making it difficult to achieve efficient sharing and utilization.

[0004] The cloud computing-based medical big data sharing platform disclosed in the present application comprises: a network construction unit for establishing a distributed storage network on a cloud computing center and connecting multiple medical institutions as independent nodes to the distributed storage network; an interface setting unit for setting two access interfaces on the cloud computing center, namely an institutional access interface and a personal access interface; a database forming unit for receiving and storing multi-dimensional medical data uploaded by multiple medical institutions based on the distributed storage network to form a medical big database; an institutional request processing unit for filtering an initial medical data set in the medical big database according to an institutional access request when the cloud computing center receives the institutional access request from the institutional access interface, the institutional access request being sent by a requesting institution; an institutional data processing unit for processing the initial medical data set based on the institutional access request through a dynamic data structure module to generate a target medical data set and sending it to the requesting institution; a personal request processing unit for filtering an associated medical data set in the medical big database according to a personal access request when the cloud computing center receives the personal access request from the personal access interface, the personal access request being sent by a requesting individual; and a personal report generating unit for reading the personal medical archives of the requesting individual, processing the associated medical data set based on the personal access request and the personal medical archives through a dynamic information conversion module to generate a personalized medical data analysis report and sending it to the requesting individual.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages: By employing a distributed storage network built on the cloud computing center through a network construction unit, multiple medical institutions are connected to this network as independent nodes, aggregating medical data scattered across various medical institutions to the cloud for centralized storage and management. An interface setting unit configures institutional and individual access interfaces within the cloud computing center, enabling differentiated services for different types of users. A database formation unit receives and stores multi-dimensional medical data uploaded by multiple medical institutions based on the distributed storage network, forming a large medical database. This aggregates and integrates the scattered medical data, creating a comprehensive and rich medical big data resource library. Through institutional request processing and institutional data processing units, when the cloud computing center receives an institutional access request, it filters the initial medical dataset from the large medical database based on the request content and processes it through a dynamic data structure module. The initial medical dataset is processed to generate a target medical dataset that meets the needs of the requesting institution and is returned, realizing personalized data services for medical institutions. Through the personal request processing unit and the personal report generation unit, when the cloud computing center receives an individual access request, it filters the medical datasets related to the individual from the medical big data database according to the request content, and combines them with the individual's health record. Through the dynamic information conversion module, it generates a personalized medical data analysis report for the user and returns it, realizing a technical solution for personalized data services for individual users. This solves the technical problem of medical data being scattered and stored in various medical institutions, making it difficult to achieve efficient sharing and utilization. It achieves the technical effect of efficient sharing and utilization of medical data by integrating and storing scattered medical data and providing data access interfaces and processing services for institutions and individuals.

[0006] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0007] Figure 1 This application provides a schematic diagram of the structure of a cloud-based medical big data sharing platform; Figure 2 This application provides a schematic diagram of a process for generating personalized medical data analysis reports in a cloud-based medical big data sharing platform.

[0008] Explanation of reference numerals in the attached figures: Network construction unit 11, Interface setting unit 12, Database formation unit 13, Organizational request processing unit 14, Organizational data processing unit 15, Individual request processing unit 16, Individual report generation unit 17. Detailed Implementation

[0009] The overall concept of the technical solution provided in this application is as follows: This application provides a cloud-based pharmaceutical big data sharing platform. First, a distributed storage network is built in a cloud computing center, connecting multiple medical institutions to achieve centralized storage and aggregation of pharmaceutical data in the cloud. Second, by setting up different access interfaces for institutions and individuals in the cloud computing center and establishing corresponding data processing modules, personalized pharmaceutical data services tailored to different user needs are provided. Finally, customized pharmaceutical datasets provided through the institutional access interface and personalized pharmaceutical data analysis reports provided through the individual access interface fully meet the different needs of medical institutions and individual users in pharmaceutical data applications.

[0010] This application constructs a cloud computing platform that integrates pharmaceutical data storage, sharing, and analysis, thereby consolidating and centrally managing scattered pharmaceutical data and providing personalized data services for different users. This effectively improves the efficiency of pharmaceutical data sharing and utilization, and can well adapt to the diverse needs of pharmaceutical big data sharing applications.

[0011] After introducing the basic principles of this application, the non-limiting embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0012] like Figure 1 As shown in the embodiment of this application, a cloud computing-based medical big data sharing platform is provided, which includes: Network building unit 11 is used to establish a distributed storage network on the cloud computing center and connect multiple medical institutions as independent nodes to the distributed storage network.

[0013] Specifically, the cloud computing center, as the core of the entire platform, provides powerful computing and storage capabilities. Network building unit 11 utilizes the infrastructure of the cloud computing center, employing virtualization technology and distributed algorithms to construct a distributed storage network on top of it. This distributed storage network adopts a decentralized architecture, not relying on any single central node, but rather consisting of multiple distributed, independent nodes, each possessing independent storage and computing resources.

[0014] After establishing the distributed storage network, network building unit 11 further connects multiple medical institutions to the network, making them independent nodes. Each medical institution, as a node, has local data storage and processing capabilities, while also synchronizing and sharing data with other nodes. By connecting medical institutions to the distributed storage network, decentralized storage and collaborative processing of medical data are achieved, improving data security, reliability, and access efficiency.

[0015] Network building unit 11 is used to build a cloud computing-based distributed storage network, integrating multiple medical institutions into the distributed storage network, providing infrastructure support for subsequent sharing and analysis of medical big data.

[0016] The interface setting unit 12 is used to set up two access interfaces in the cloud computing center, namely an institutional access interface and a personal access interface.

[0017] Specifically, in order to meet the needs of different user groups, the interface setting unit 12 sets up two types of access interfaces in the cloud computing center: an institutional access interface and a personal access interface.

[0018] The institutional access interface is an interface for medical institutions. Through this interface, medical institutions can upload their local medical data to the cloud computing center and also access and obtain shared medical data resources from other institutions. The institutional access interface requires high security and access control to ensure data confidentiality and integrity. When designing the institutional access interface, interface setting unit 12 employs technologies such as authentication and encrypted transmission, and implements fine-grained access control based on the roles and permissions of different institutions.

[0019] The personal access interface is designed for ordinary users, especially individual patients. Through this interface, users can query and access their own medical and health data, such as electronic medical records, test reports, and medication records. Simultaneously, the personal access interface provides users with personalized health management services, such as disease risk assessment, medical guidance, and health education. Compared to institutional access interfaces, personal access interfaces have higher requirements for convenience and user experience. Interface setting unit 12 adopts a simple and user-friendly interface design and easy-to-understand interaction methods to lower the barrier to entry, allowing ordinary users to easily access and utilize medical big data resources.

[0020] The interface setting unit 12 provides access portals for different user groups to the medical big data sharing platform, meeting the differentiated needs of institutional and individual users in terms of data sharing, data acquisition, and personalized services, and promoting the efficient utilization and sharing of medical data.

[0021] Database forming unit 13 is used to receive and store multi-dimensional medical data uploaded by the multiple medical institutions based on the distributed storage network, forming a large medical database.

[0022] Specifically, when medical institutions upload pharmaceutical data through the institutional access interface, database formation unit 13 receives and processes the uploaded data. First, it performs format validation and integrity checks on the data to ensure its standardization and accuracy. Then, based on the data's characteristics and attributes, it categorizes and stores it on different nodes of a distributed storage network. Distributed storage not only improves data storage capacity and access efficiency but also enhances data security and reliability, preventing data loss due to single points of failure. During data storage, database formation unit 13 organizes and manages the pharmaceutical data in multiple dimensions, extracting key attributes such as disease type, drug name, and patient characteristics, and establishing indexes and relationships based on these attributes. This allows for rapid location and retrieval of required data in subsequent data retrieval and analysis, enabling multi-dimensional and multi-faceted data mining and utilization. As medical institutions continuously upload various types of pharmaceutical data, database formation unit 13 continuously expands and updates, forming a large pharmaceutical database. Simultaneously, it regularly cleans, deduplicates, and integrates the data to ensure the database's data quality and consistency.

[0023] Database forming unit 13 constructs a comprehensive and high-quality pharmaceutical big data database, providing rich and reliable data resources for medical institutions and individual users. Through the receipt, storage, and management of multi-dimensional pharmaceutical data, database forming unit 13 lays a solid data foundation for the sharing, analysis, and application of pharmaceutical big data.

[0024] The organization request processing unit 14 is used to filter the initial medical dataset in the medical big data database according to the organization access request when the cloud computing center receives the organization access request from the organization access interface. The organization access request is sent by the requesting organization.

[0025] Specifically, medical institutions can initiate data access requests to the cloud computing center through the institutional access interface according to their own business needs. The institutional request processing unit 14 is specifically responsible for processing and responding to these institutional access requests. Upon receiving an institutional access request, the institutional request processing unit 14 first verifies the legality and permissions of the request to ensure that the requesting institution has the corresponding data access permissions. After verification, the institutional request processing unit 14 parses the content of the institutional access request, extracting key information such as the data type, time range, and geographical region required by the requesting institution, forming structured query conditions. Next, based on these query conditions, the institutional request processing unit 14 performs data filtering in the medical big data database, utilizing the database's indexing mechanism and query optimization techniques to quickly locate and retrieve medical data records that match the query conditions. These filtered data records constitute the initial medical dataset, serving as the foundation for subsequent data processing and analysis.

[0026] During the initial screening of the medical dataset, the institution request processing unit 14 considers both data privacy protection and secure access control. Based on the requesting institution's role and permissions, it performs necessary anonymization processing on the screened data, such as removing patient identity information and hiding sensitive data, to ensure privacy compliance during data sharing. Simultaneously, the institution request processing unit 14 optimizes the data volume and transmission method of the initial medical dataset to improve data acquisition efficiency and performance.

[0027] The institution request processing unit 14 responds to and processes data access requests from requesting institutions. Based on the specific needs of the institution's access request, it quickly and accurately filters out the required initial medical dataset from the medical big data database, providing the requesting institution with a convenient and efficient data acquisition channel and promoting data sharing and collaboration among medical institutions.

[0028] The institutional data processing unit 15 is used to process the initial medical dataset based on the institutional access request through the dynamic data structure module, generate a target medical dataset, and send it to the requesting institution.

[0029] Specifically, although the initial medical dataset has undergone preliminary screening based on the institution's access request, its data format and content structure do not yet directly meet the specific application needs of the requesting institution. Therefore, the institution data processing unit 15 introduces a dynamic data structure module to perform in-depth processing and transformation on the initial medical dataset, generating a targeted and easy-to-use target medical dataset.

[0030] First, the institutional data processing unit 15 conducts an in-depth analysis of the institutional access request, extracting the requesting institution's data processing requirements regarding data format, content organization, and feature representation. Then, these requirements are passed to the dynamic data structure module, guiding it to perform structured processing on the initial medical dataset. The dynamic data structure module employs a flexible and configurable data representation and transformation mechanism, dynamically adjusting the data organization and presentation format according to different data processing needs. For example, it parses and splits the initial medical dataset, extracting key fields and attributes, and performing field filtering, data cleaning, and format conversion operations according to data processing requirements. Through these processes, the dynamic data structure module transforms the initial medical dataset into a structured and standardized target medical dataset, enabling its direct application in the requesting institution's business systems and analytical tools. During processing, the institutional data processing unit 15 verifies the quality and integrity of the target medical dataset to ensure it meets the requesting institution's data quality requirements. Simultaneously, considering data transmission efficiency and security, it selects appropriate data compression and encryption algorithms before sending the target medical dataset to the requesting institution.

[0031] Through the institutional data processing unit 15, the initial medical dataset is processed and customized using a dynamic data structure module to generate a target medical dataset that meets the specific needs of the requesting institution. This fully considers the differentiated needs of different institutions in data application and provides a flexible and efficient data processing and transformation mechanism, enabling the requesting institution to quickly obtain usable, high-quality medical data to support its business operations and scientific research activities.

[0032] The personal request processing unit 16 is used to filter related medical datasets in the medical big data database according to the personal access request when the cloud computing center receives a personal access request from the personal access interface. The personal access request is sent by the requesting individual.

[0033] Specifically, besides medical institutions, individual users, especially patients, also have a need to access and query medical and health data. The personal request processing unit 16 is responsible for processing and responding to data access requests from individual users, providing them with convenient and secure personal medical data query services. In the pharmaceutical big data sharing platform, individual users can not only query their own medical and health data, but also obtain other related pharmaceutical information and knowledge based on their own medical information and query needs. The personal request processing unit 16 provides personalized and comprehensive pharmaceutical information query and sharing services to individual users through intelligent parsing and correlation analysis of personal access requests.

[0034] Upon receiving a personal access request from a requesting individual, the personal request processing unit 16 first authenticates the individual's identity to ensure they have legitimate data access rights. Then, it conducts in-depth analysis of the request, extracting two key types of information: first, the individual's personal medical information, such as past medical history, current symptoms, and test results; and second, the individual's query content, such as desired disease knowledge, treatment plans, and medication guidance. Based on these two types of key information, the personal request processing unit 16 performs data filtering and correlation analysis within a large medical database. On one hand, it retrieves the individual's historical medical data, using it as the starting point and background for analysis; on the other hand, based on the individual's query content, it utilizes medical ontology, semantic analysis, and other technologies to discover and match relevant medical information within the large medical database, such as clinical guidelines, research literature, drug instructions, and case reports. Through correlation data mining based on specific individual needs, the personal request processing unit 16 filters out relevant and targeted medical datasets closely related to the requesting individual from the large medical database.

[0035] The personal request processing unit 16 responds to and processes individual users' medical information query requests. Based on the individual's own medical information and query needs, it dynamically filters and associates relevant medical datasets in the medical big data database, thus meeting the individual users' needs for information acquisition and sharing in the field of medical and health care.

[0036] The personal report generation unit 17 is used to read the personal medical records of the requesting individual, and through the dynamic information conversion module, process the associated medical dataset based on the personal access request and the personal medical records to generate a personalized medical data analysis report, which is then sent to the requesting individual.

[0037] Specifically, in the medical big data sharing platform, in order to enable individual users to better understand and apply the medical data obtained from the platform, the personal report generation unit 17 comprehensively utilizes personal medical records and related medical datasets, and through the processing of the dynamic information conversion module, generates personalized medical data analysis reports that meet the characteristics and needs of individual users.

[0038] After receiving the relevant medical dataset filtered by the personal request processing unit 16, the personal report generation unit 17 first reads the requester's personal medical record. The personal medical record contains the requester's basic information, past medical history, lifestyle habits, and other information, serving as a crucial basis for personalized data analysis. Next, the personal report generation unit 17 passes the personal medical record and the relevant medical dataset to the dynamic information conversion module for processing. The dynamic information conversion module employs adaptive data interpretation and conversion algorithms, dynamically adjusting the representation and presentation strategies of medical data based on the characteristics of different individual users. During processing, the dynamic information conversion module deeply analyzes the personal medical record, extracting key information such as the requester's health status, disease risk, and medication history, and comparing and integrating this information with the knowledge in the relevant medical dataset. Through personalized information matching and integration, the dynamic information conversion module can identify the most relevant and valuable medical data content for the requester and provide targeted interpretation and explanation. Building upon the information conversion, the dynamic information conversion module further considers factors such as the individual user's cognitive level and reading habits. Using easily understandable language and intuitive charts, it transforms the pharmaceutical data analysis results into personalized pharmaceutical data analysis reports. These reports not only include the pharmaceutical information requested by the individual but also provide personalized health management suggestions and medical guidance, helping individual users better understand and utilize pharmaceutical big data. Subsequently, the personal report generation unit 17 sends the generated personalized pharmaceutical data analysis report to the requesting individual, realizing the shared application of pharmaceutical big data in personal health management.

[0039] Through the personal report generation unit 17, based on personal medical records and associated pharmaceutical datasets, and through the intelligent processing of the dynamic information conversion module, a personalized pharmaceutical data analysis report is generated for individual users. It fully considers factors such as the individual user's health status and cognitive characteristics, and provides tailored pharmaceutical information interpretation and health management services, enabling individual users to conveniently, efficiently and accurately obtain and apply pharmaceutical big data, and realize personalized sharing of pharmaceutical big data.

[0040] Furthermore, embodiments of this application also include: Parse the organization access request and extract the request parameters, which include data type, time range, and geographical range; Verify the data access permissions of the requesting organization and establish a retrieval index in the medical database based on the request parameters; Based on the search index and the data access permissions, data that meets the conditions is filtered in the medical database, and the filtered data is de-identified to form the initial medical dataset.

[0041] In one feasible implementation, the institution request processing unit 14 first parses the received institution access request, extracting request parameters, including data type, time range, and geographical range, to specify the characteristics and scope of the medical data required by the requesting institution. Specifically, the data type specifies the type of medical data the requesting institution is interested in, such as electronic medical records, medical orders, and test reports; the time range specifies the start and end dates of the data required by the requesting institution; and the geographical range specifies the geographical area of ​​the data required by the requesting institution, such as a specific hospital, city, or province. By extracting the request parameters, the institution request processing unit 14 can clarify the specific data needs of the requesting institution, providing a basis for subsequent data filtering. After extracting the request parameters, the institution request processing unit 14 verifies the data access permissions of the requesting institution. Different medical institutions have different levels of data access permissions; for example, some institutions can only access data of a specific type or time range, while others have broader data access permissions. The institution request processing unit 14 verifies the identity and permissions of the requesting institution to ensure that it has legitimate and appropriate data access permissions. After verification, the institution request processing unit 14 establishes a corresponding search index in the medical database based on the request parameters. A retrieval index is a data structure created to accelerate data retrieval. It organizes and indexes data in a large medical database based on conditions specified in the request parameters, such as data type, time range, and geographical scope. By establishing a retrieval index, the efficiency and accuracy of data filtering are improved.

[0042] After establishing the search index, the institutional request processing unit 14 efficiently filters data that meets the request criteria from the pharmaceutical database. Simultaneously considering the request parameters and the requesting institution's data access permissions, only pharmaceutical data that the requesting institution has the right to access and that meets the request parameter requirements is selected. In this way, the filtered data not only meets the actual needs of the requesting institution but also takes into account data security and privacy protection. To further protect patient privacy, the institutional request processing unit 14 performs anonymization processing on the filtered data, removing or replacing sensitive information such as patient names, ID numbers, and home addresses, making this information unassociatable with any specific individual. The anonymized pharmaceutical data retains the original data structure and key content while effectively protecting patient privacy. After filtering and anonymization, the pharmaceutical data that meets the request criteria and privacy protection requirements forms the initial pharmaceutical dataset, containing the preliminary scope and content of the pharmaceutical data required by the requesting institution, laying the foundation for subsequent data processing and analysis.

[0043] By parsing institutional access requests, verifying data access permissions, building search indexes, filtering data that meets the criteria, and performing anonymization processing, an initial medical dataset that meets the needs and privacy protection requirements of the requesting institution is selected from the medical big data database. This ensures the relevance, security, and compliance of medical data sharing, providing requesting institutions with high-quality and usable medical data resources to support their various medical-related research and applications.

[0044] Furthermore, embodiments of this application also include: The initial medical dataset is deconstructed into multiple data segments, and the structural and semantic features of each data segment are extracted. Extract the processing requirements of the institution's access request, generate fragment combination rules based on the processing requirements, the structural features, and the semantic features, and reorganize the multiple data fragments according to the fragment combination rules to form multiple initial medical data subsets; The relevance of multiple initial medical data subsets to the institution's access request is evaluated, and the fragment combination rules are iteratively adjusted based on the relevance evaluation results to dynamically optimize the recombination process and form multiple optimized medical data subsets. Multiple optimized medical data subsets are integrated according to semantic hierarchy to eliminate semantic and format differences between the multiple optimized medical data subsets, forming the target medical dataset.

[0045] In a preferred embodiment, firstly, the dynamic data structure module in the institutional data processing unit 15 deconstructs the initial medical dataset, splitting it into multiple data segments. The deconstruction process is based on the inherent structure and semantics of medical data, such as dividing it according to data type, time period, patient group, etc. Each deconstructed data segment is a subset of the initial medical dataset, containing some of the original data. Simultaneously with the deconstruction, the dynamic data structure module extracts the structural and semantic features of each data segment. Structural features describe information such as the data format, field composition, primary key and foreign key relationships of the data segment, reflecting the structural attributes of the segment. Semantic features describe the medical concepts, clinical significance, medication events, etc., contained in the data segment, reflecting the semantic attributes of the data segment. Extracting the structural and semantic features of each data segment provides a basis for subsequent data segment reassembly, ensuring that the reassembled data subset has structural and semantic integrity and consistency.

[0046] After extracting the structural and semantic features of each data fragment, the dynamic data structure module further analyzes the institution's access requests, extracting the implicit processing requirements, including operations such as data aggregation and statistics, correlation analysis, and data mining, reflecting the requesting institution's specific application needs for medical data. Based on the processing requirements and the structural and semantic features of the data fragments, the dynamic data structure module generates fragment combination rules. These rules define how to select and combine different data fragments to meet processing requirements and ensure structural and semantic consistency. On one hand, the rules consider structural compatibility between data fragments, such as data type matching and primary / foreign key relationships, ensuring the structural integrity of the combined data. On the other hand, they consider semantic relevance between fragments, such as associations of medical concepts and the chronological order of clinical events, ensuring semantic coherence of the combined data. According to the fragment combination rules, the dynamic data structure module reorganizes multiple data fragments to form multiple initial medical data subsets. Each initial medical data subset is a combination of several data fragments, containing the data content necessary to meet processing requirements, and achieving organic integration between fragments in terms of structure and semantics. During the recombination process, a data fragment may be used multiple times, participating in the formation of multiple initial medical data subsets. Simultaneously, some data fragments may not be used because they do not meet the combination rules. Therefore, the relationship between data fragments and initial medical data subsets is many-to-many.

[0047] After forming multiple initial medical data subsets, the dynamic data structure module evaluates the relevance of each subset to the institution's access request. The relevance assessment examines the data content, structure, and semantics of the initial medical data subsets to determine their degree of matching with the requesting institution's processing needs. The assessment process involves multiple dimensions, including data quality, statistical indicators, and keyword matching, to quantify the relevance of the subsets to the institution's access request, thus obtaining the relevance assessment results. Subsequently, based on the relevance assessment results, the dynamic data structure module iteratively adjusts the fragment combination rules. If the assessment finds that some subsets have low relevance to the access request, it indicates that the current fragment combination rules are insufficient and need optimization. First, the reasons for insufficient relevance are analyzed, such as missing data fragments or semantic inconsistencies, and then the combination rules are modified accordingly, such as supplementing missing data fragments or adjusting the fragment combination method. Through iterative optimization, the dynamic data structure module continuously improves the matching degree between the initial medical data subsets and the institution's access request, generating a dataset that better meets the needs of the requesting institution. After multiple rounds of iterative optimization, the dynamic data structure module will form several optimized medical data subsets. These subsets are not only highly relevant to the access requests in terms of data content, but also achieve good coordination and consistency in terms of structure and semantics, which can effectively support the data processing and analysis work of the requesting institutions.

[0048] After obtaining optimized medical data subsets, the dynamic data structure module further integrates these subsets semantically. The purpose of semantic integration is to eliminate potential semantic and format differences between different subsets, enabling them to form a unified and coherent whole semantically. The integration process first extracts semantic information from each optimized medical data subset, such as medical concepts, event times, and patient characteristics. Then, this semantic information is compared and mapped to identify differences in semantic representation between different subsets, such as different names for the same concept or different time formats for the same event. Through semantic mapping and transformation, the dynamic data structure module eliminates these semantic differences, enabling different subsets to achieve semantic unification and standardization. Based on semantic unification, the dynamic data structure module performs necessary conversions and alignments on the data formats of the optimized medical data subsets, eliminating format differences between subsets, such as inconsistencies in field naming, data types, and encoding methods. Through format unification, different subsets achieve seamless integration at the physical storage and access levels. Finally, the dynamic data structure module integrates the semantically and format-unified optimized medical data subsets to form the target medical dataset. The target medical dataset is a structured and semantically transformed version of the initial medical dataset. It not only retains the core content of the original data, but also achieves a high degree of consistency in semantics and format, making it easier for requesting institutions to perform unified data processing and analysis, and fully leveraging the value of medical big data.

[0049] Through the dynamic data structure module, the initial medical dataset is subjected to semantic-driven structured processing to generate a semantically coherent and structurally coordinated target medical dataset. It fully considers the processing needs of institutional access requests. By utilizing the structural and semantic features of data fragments, and through dynamic fragment combination, optimization, and integration processes, the original medical data is reorganized into a high-quality and easy-to-use target dataset, thereby effectively supporting medical institutions in carrying out the application of medical data.

[0050] Furthermore, embodiments of this application also include: Extract semantic information from the data in the multiple optimized medical data subsets to construct a medical semantic vector space; Cluster analysis is performed in the pharmaceutical semantic vector space to identify semantic core regions. The multiple optimized pharmaceutical data subsets are mapped to the most relevant semantic core regions to achieve data subset classification based on core semantics. Within each semantic core region, the semantic representation of a data subset is unified, the data format is adjusted, semantic and format differences are eliminated, and then the data of all semantic core regions are associated according to preset semantic linking rules to form the target medical dataset.

[0051] In a preferred embodiment, firstly, semantic analysis is performed on each optimized medical data subset to extract the semantic information it contains, including multiple aspects such as medical concepts, clinical events, drug treatments, and patient characteristics, reflecting the semantic content of the optimized medical data subset. For example, through technologies such as medical ontology, semantic dictionaries, and named entity recognition, unstructured or semi-structured data such as text and encoding of the optimized medical data subset are transformed into structured semantic representations. Based on the extracted semantic information, a dynamic data structure module constructs a medical semantic vector space. This space is a high-dimensional mathematical model, where each dimension represents a semantic feature in the medical field, such as disease type, drug effect, and treatment plan. By mapping the extracted semantic information to the semantic vector space, each optimized medical data subset can be represented by a semantic vector, and the value of each component of the semantic vector reflects the weight or correlation of the optimized medical data subset on the corresponding semantic feature.

[0052] After constructing the medical semantic vector space, cluster analysis is performed within this space to group semantically similar and content-related data subsets into several semantic core regions. Each semantic core region represents a semantic theme or concept, such as a certain disease or a certain type of drug, reflecting the common semantic features of a set of data subsets. Each semantic core region contains a cluster of semantically related optimized medical data subsets, and each optimized medical data subset is mapped to the core region most relevant to its semantic vector. Through cluster analysis and mapping, data subset classification based on core semantics is achieved. The classified optimized medical data subsets are more semantically focused and consistent. Subsets within the same semantic core region have high semantic homogeneity, while subsets between different semantic core regions have significant semantic differences. After completing the division of the semantic core regions, the semantic representation of the optimized medical data subsets within each region is unified and the format is adjusted. Although subsets within the same region are semantically related, there may still be some differences in semantic representation and data format, such as inconsistencies in synonyms, abbreviations, and units. To eliminate these differences, referencing medical ontology and semantic mapping rules, the semantic representations of subsets within a region are unified to standardized medical terminology and concepts, ensuring consistency and standardization of semantic expression. Simultaneously, necessary data format conversions and alignments are performed on the subsets to ensure consistency in data types, field naming, and encoding methods, facilitating subsequent data integration and processing. Based on eliminating semantic and format differences, data subsets from different semantic core regions are associated and integrated according to pre-defined semantic linking rules. These rules define the logical relationships and order between different semantic core regions, such as the correspondence between diseases and drugs, and the order of examinations and diagnoses. Through semantic linking, optimized medical data subsets from different regions form an organic whole at the semantic level, reflecting a complete clinical diagnosis and treatment process and medical logic. The data subsets from multiple semantic core regions after semantic linking are integrated into a target medical dataset. This target medical dataset achieves unified and standardized semantic representation, consistency and compatibility in data format, and also reflects the semantic logical relationships between different data subsets, forming a semantically coherent and structurally complete medical dataset, facilitating data analysis and application for requesting institutions.

[0053] By constructing a pharmaceutical semantic vector space and performing cluster analysis within it, the semantic classification of optimized pharmaceutical data subsets and the identification of semantic core regions are achieved. Based on this, the semantic representation of the optimized pharmaceutical data subsets within each semantic core region is further unified and the format adjusted. Subsets from different regions are then linked and integrated based on semantic linking rules to form a semantically coherent and format-consistent target pharmaceutical dataset. This achieves hierarchical semantic organization and integration of data subsets, improves the semantic interoperability and integration quality of pharmaceutical data, and provides better support for the efficient sharing and utilization of pharmaceutical big data.

[0054] Furthermore, embodiments of this application also include: Parse the personal access request to extract the personal identification, request content, and request parameters; Based on the personal identification, retrieve the medical records related to the requested individual from the medical database to obtain the individual's medical file; Perform semantic analysis on the request content, and construct a medical data retrieval tree based on the request parameters; Traverse the medical data retrieval tree. At non-leaf nodes, dynamically generate medical data filtering rules based on the personal medical records. At leaf nodes, use the corresponding medical data filtering rules to retrieve relevant data from the medical database to form a preliminary personal dataset. The correlation between each data point in the preliminary personal dataset and the personal medical record is evaluated, and data with a correlation below a preset threshold is removed to obtain the associated medical dataset.

[0055] In one feasible implementation, firstly, the personal request processing unit 16 parses the received personal access request, extracting key information including a personal identifier, request content, and request parameters. The personal identifier uniquely identifies the individual user initiating the request; the request content indicates the subject or category of the medical data required by the individual user, such as disease diagnosis, medication guidance, or health management; and the request parameters further clarify the specific requirements of the request content, such as the data time range, hospital scope, and severity of the condition. By parsing the personal access request, the personal request processing unit 16 accurately understands the individual user's data needs, providing a basis for subsequent data filtering. After extracting the personal identifier, the personal request processing unit 16 uses this identifier to retrieve medical records related to the requesting individual from the medical database. The medical database stores medical data from multiple medical institutions, including individual medical records. Individual medical records summarize an individual's medical information from different medical institutions and at different times, such as past medical history, allergy history, physical examination reports, and imaging data, reflecting the individual's overall health status and medical background. By associating with the identity identifier, the personal request processing unit 16 accurately locates the medical record corresponding to the requesting individual in the medical big data database, laying the foundation for subsequent personalized data screening.

[0056] After obtaining the individual's medical records, the personal request processing unit 16 performs semantic analysis on the request content. Utilizing natural language processing technology and combining it with medical knowledge, it achieves a deeper understanding and interpretation of the request content. Through semantic analysis, it clarifies the semantic elements involved in the request content, such as medical concepts, disease types, and treatment methods, and extracts keywords and semantic relationships. Based on the semantic analysis, the personal request processing unit 16 constructs a medical data retrieval tree in conjunction with the request parameters. The medical data retrieval tree is a hierarchical data organization structure. It uses the semantic elements of the request content as its core and the request parameters as constraints, forming a top-down, coarse-to-fine retrieval path. The root node of the medical data retrieval tree represents the topic of the request content, such as "diabetes," intermediate nodes represent various levels of semantic subdivision, such as "drug treatment" and "dietary control," and leaf nodes represent the finest-grained data requirements, associated with corresponding request parameters, such as "prescriptions for hypoglycemic drugs in the past year." After constructing the medical data retrieval tree, the personal request processing unit 16 traverses the medical data retrieval tree. During the traversal, for non-leaf nodes of the medical data retrieval tree, the personal request processing unit 16 dynamically generates medical data filtering rules based on the individual's medical records. These filtering rules are a set of conditional expressions used to select data from the medical database that is relevant to the individual's situation and matches the search topic. The generation of these filtering rules considers key information in the individual's medical records, such as age, gender, and medical history, ensuring that the filtered data matches the individual's actual situation, thus achieving personalized and precise data filtering. For leaf nodes of the retrieval tree, the personal request processing unit 16 directly uses the corresponding medical data filtering rule to retrieve relevant data from the medical database. The medical data filtering rules, as query conditions, together with the request parameters, determine the specific range and characteristics of the required data. By executing the medical data filtering rules in the large database, the personal request processing unit 16 quickly and efficiently obtains raw medical data related to the request content and the individual's situation, and combines it to form a preliminary personal dataset containing the basic data content required by the request. After forming the preliminary personal dataset, the personal request processing unit 16 further evaluates the relevance of each data point to the individual's medical records. For example, by calculating the matching degree between data and personal medical records across multiple dimensions such as time, disease type, and severity, the degree of correlation between the two can be quantified. A preset relevance threshold is set and used to filter the initial personal dataset, removing data with relevance below the threshold and retaining highly relevant and well-matched data. After relevance assessment and filtering, the initial personal dataset is further optimized and streamlined to form a related medical dataset. This related medical dataset, tailored to the specific needs and circumstances of individual users, covers various types of closely related medical data, such as disease diagnosis and treatment information, medication history, and lifestyle guidance, providing comprehensive, accurate, and personalized medical information services.

[0057] Furthermore, such as Figure 2 As shown, embodiments of this application also include: Extract the points of interest from the personal access request and the health indicators from the personal medical record; Based on the points of interest and the health indicators, the related medical datasets are classified and sorted to form multiple thematic data subsets; For each thematic data subset, statistical analysis and trend prediction are performed in conjunction with historical data from the individual's medical records to generate corresponding thematic analysis results; The results of multiple thematic analyses are integrated to generate the personalized medical data analysis report, which includes personalized recommendations and risk warnings.

[0058] In one feasible implementation, the personal report generation unit 17 first analyzes personal access requests and personal medical records, extracting key information. For personal access requests, it extracts content reflecting the user's concerns, such as disease types, medication types, and health management aspects, reflecting the user's subjective intentions and interests, and serving as an important basis for generating personalized reports. For personal medical records, it extracts objective indicators reflecting the user's health status, such as physiological parameters, test results, and medical history, depicting the individual's physical condition and medical characteristics from different perspectives, and providing key data for personalized analysis and prediction. After extracting the concerns and health indicators, the personal report generation unit 17 uses this as a basis to classify and sort the associated medical dataset. The classification process divides the associated medical dataset into multiple thematic data subsets based on the concerns and health indicators. Each thematic data subset focuses on a specific health theme or medical need, such as chronic disease management, medication safety, and health risk assessment, which are closely related to the user's concerns and health status. Based on the classification, the personal report generation unit 17 sorts the data within each subject data subset and determines the priority and display order of the data according to factors such as the time, importance, and reliability of the data.

[0059] After forming thematic data subsets, the personal report generation unit 17 conducts in-depth analysis of each subset. The analysis process fully utilizes historical data from the individual's medical records, such as past examination results, medical records, and lifestyle information, comparing and correlating this data with current data within the thematic data subsets. Through statistical analysis, it calculates the average values, trends, and anomalies of various health indicators, assessing the individual's health status and medical needs under that theme. Based on the statistical analysis, the personal report generation unit 17 applies predictive models to predict the individual's health trends over a future period. The predictive model comprehensively considers factors such as the individual's historical data, current condition, and theme characteristics to estimate the potential health risks or room for improvement the individual may face under that theme, providing a reference for personalized decision-making. The results of statistical analysis and trend prediction form thematic analysis results, comprehensively reflecting the individual's current status, trends, and potential risks under a specific health theme, which is the core content of the personalized medical data analysis report. After generating the analysis results for each theme, the personal report generation unit integrates them to form a complete personalized medical data analysis report. Based on the thematic analysis results, the personal report generation unit 17 provides personalized health management suggestions and risk warnings, offering practical action guidance and alerts to individual users. For example, regarding the topic of chronic disease management, it is recommended that individuals undergo regular check-ups, adjust their dietary habits, and maintain regular exercise; regarding the topic of medication safety, individuals are advised to pay attention to the adverse reactions of certain medications, avoid adjusting dosages on their own, and have regular follow-up visits.

[0060] Furthermore, embodiments of this application also include: A personal health model is established, which is constructed based on the individual's medical records and a medical knowledge base; The personal health model is used to assign weights to the data in the subject data subset. Based on the weight assignment results, time series analysis is performed on the subject data subset to obtain time series analysis results, including the changing trends of key health indicators. The changing trends of key health indicators are compared with the expected trends in the personal health model to generate deviation analysis results. Based on the deviation analysis results and the aforementioned medical knowledge base, a personalized health risk assessment is generated. The time series analysis results, the bias analysis results, and the health risk assessment results are integrated to form the thematic analysis results.

[0061] In a preferred embodiment, firstly, a personal health model is established using two data sources: personal medical records and a medical knowledge base, to describe and predict an individual's health status. The personal health model is comprehensive. Personal medical records provide historical health data, such as physiological indicators, diagnostic results, and treatment plans, reflecting the objective manifestation of an individual's health status. The medical knowledge base provides general principles and professional knowledge in the medical field, such as disease mechanisms, prognostic factors, and treatment guidelines, providing a theoretical basis for the analysis and prediction of an individual's health status. Through machine learning algorithms, the data from the personal medical records and the knowledge from the medical knowledge base are integrated and trained to construct a personal health model that accurately reflects the individual's health characteristics. After establishing the personal health model, the personal report generation unit 17 uses this model to analyze each subject data subset. First, the personal health model assigns weights to each data item in the subject data subset. The weights reflect the degree of influence of different data on an individual's health status; for example, abnormalities in certain physiological indicators may be more indicative of health problems than other indicators. Based on medical knowledge and individual characteristics, the personal health model calculates the weights of each data item, highlighting key factors and reducing the influence of secondary factors. Based on the weighted allocation, the personal report generation unit 17 performs time series analysis on the thematic data subset. Time series analysis focuses on the changes in data over time to discover the evolution trends and patterns of an individual's health status. For each key health indicator, its changes over a period of time are analyzed, such as increases, decreases, and fluctuations, to determine its overall trend and any abnormalities. The results of the time series analysis form the changing trends of key health indicators, intuitively displaying the dynamic changes in an individual's health status.

[0062] After obtaining the changing trends of key health indicators, these trends are compared with the expected trends in the personal health model. The expected trends are estimates of the normal range and direction of change of personal health indicators based on factors such as basic personal information, lifestyle, and genetic background, combined with medical knowledge. By comparing the actual and expected trends, abnormal deviations and potential risks in an individual's health status are identified. Subsequently, the degree of deviation between the actual and expected trends is calculated, generating deviation analysis results that quantitatively describe the gap between an individual's actual and expected health status, indicating the direction and magnitude of the deviation. Positive deviations indicate that an individual's health status is better than expected, while negative deviations suggest potential health risks. Afterwards, the personal report generation unit 17 matches and infers from the deviation analysis results with epidemiological data and clinical guidelines in the medical knowledge base to estimate the probability and severity of occurrence of different health risks, such as chronic disease risk and adverse drug reaction risk. During the assessment process, individual specific circumstances, such as age, gender, and past medical history, are fully considered to provide personalized risk levels and explanations. Personalized health risk assessments can alert individuals to specific health problems and enable them to take targeted prevention and management measures. After generating time series analysis results, bias analysis results, and health risk assessments, the individual report generation unit 17 integrates them to form complete thematic analysis results, comprehensively showcasing an individual's health status, trends, abnormalities, and potential risks under a specific health theme, covering multiple analytical dimensions of descriptive and predictive nature.

[0063] Furthermore, embodiments of this application also include: Based on the preset report template and the analysis results of the multiple topics, an initial report structure is generated, which contains multiple report chapters; Information is extracted from the analysis results of the multiple topics and converted into text descriptions, which are then filled into the corresponding report sections. Based on the user characteristics and preferences in the personal medical records, the populated report sections are personalized and sorted to form the personalized medical data analysis report.

[0064] In one feasible implementation, the report template is a general report structure framework that predefines the basic components, chapter divisions, and layout styles of the report, providing a unified and standardized starting point for report generation. The personal report generation unit 17 matches and maps multiple thematic analysis results to the report template, identifying the corresponding report chapters for each thematic analysis result, including basic personal information, health overview, analysis of major health problems, trend prediction and risk warnings, and lifestyle recommendations. Each chapter has its predetermined content scope and presentation requirements. By mapping the thematic analysis results to the corresponding report chapters, a complete and content-rich initial report structure is generated. After generating the initial report structure, natural language generation technology is applied to extract key information from the thematic analysis results and transform it into easily understandable text descriptions. For example, for time series analysis results, the changing trends of key health indicators are extracted, generating text descriptions such as "Over the past year, your blood pressure level has generally shown an upward trend, especially with a significant increase in the last three months." Based on the health risk assessment results, the risk level and main influencing factors are extracted, generating a text description similar to, "Based on your age, BMI, and family medical history, you have a high risk of developing Type-2 diabetes within the next five years. Regular blood glucose monitoring and a reasonable diet are recommended." This text description is then inserted into the corresponding sections of the initial report structure, making the report more detailed, vivid, and easy to understand. After the report content is filled in, the personal report generation unit 17 personalizes the report sections based on user characteristics and preferences recorded in the individual's medical records, forming the final personalized medical data analysis report. User characteristics and preferences reflect the individual user's focus, comprehension ability, and acceptance habits in health management, such as age group, education level, past medical history, and lifestyle. The personal report generation unit 17 analyzes user characteristics and preferences, adjusting the order and focus of the report sections to make the overall report structure more aligned with the user's actual needs and acceptance habits. For example, for elderly users, priority is given to displaying content on chronic disease management and medication safety, presented in a more concise and visually appealing manner. For users focused on health and wellness, the emphasis is on lifestyle improvements and health risk prevention, with more background information and operational guidance provided. Personalized sorting allows the report's structure and focus to dynamically adapt to the characteristics of different users, improving the report's relevance and practicality. After personalized sorting, a well-structured, content-rich, accurate, and focused personalized medical data analysis report is output, comprehensively presenting the individual user's health status, medical needs, and management recommendations, providing tailored health management references.

[0065] In summary, the cloud-based medical big data sharing platform provided in this application has the following technical effects: The network construction unit is used to establish a distributed storage network on the cloud computing center, connecting multiple medical institutions as independent nodes to the distributed storage network. This enables the aggregation and centralized storage of pharmaceutical data in the cloud, breaking down data barriers between medical institutions. The interface setting unit is used to set up two access interfaces in the cloud computing center: an institutional access interface and a personal access interface. This provides differentiated data access channels for different types of users, meeting their diverse needs. The database formation unit is used to receive and store multi-dimensional pharmaceutical data uploaded by multiple medical institutions based on the distributed storage network, forming a large pharmaceutical database. This integrates scattered and heterogeneous pharmaceutical data, building a comprehensive and rich pharmaceutical big data resource library, laying the foundation for subsequent data sharing and analysis. The institutional request processing unit is used when the cloud computing center receives institutional access requests from the institutional access interface. Based on the institutional access requests, it filters the initial pharmaceutical dataset in the large pharmaceutical database according to the requests. Institutional access requests are sent by the requesting institutions. Based on the specific needs of the medical institutions, it quickly locates and filters the required subset of pharmaceutical data, improving the targeting and efficiency of data acquisition. The Institutional Data Processing Unit, through a dynamic data structure module, processes initial medical datasets based on institutional access requests to generate target medical datasets, which are then sent to the requesting institution. It further processes the initial medical datasets according to the individual needs of the medical institution, generating customized medical datasets and providing high-quality data services. The Personal Request Processing Unit, when the cloud computing center receives personal access requests from the personal access interface, filters relevant medical datasets in the medical database based on the personal access request. Since personal access requests are sent by the individual, it quickly locates and filters relevant subsets of medical data based on the individual user's specific needs, improving the targeting and efficiency of data acquisition. The Personal Report Generation Unit reads the requesting individual's personal medical records and, through a dynamic information conversion module, processes the relevant medical datasets based on the personal access request and personal medical records to generate personalized medical data analysis reports, which are then sent to the requesting individual. It also generates targeted health analysis reports based on the individual user's health records and relevant medical data, providing high-value, personalized shared medical data. By integrating and storing scattered medical data and providing data access interfaces and processing services for institutions and individuals, efficient sharing and utilization of medical data are achieved.

[0066] In summary, any step of the platform described above can be stored as a computer instruction or program in an unrestricted computer memory, and can be called and identified by an unrestricted computer processor to implement any of the platforms in the embodiments of this application, without any additional restrictions.

[0067] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application intends to include such modifications and variations.

Claims

1. A cloud computing-based medical big data sharing platform, characterized in that, The method comprises the following steps: a network construction unit is configured to establish a distributed storage network on a cloud computing center, and access multiple medical institutions as independent nodes to the distributed storage network; an interface setting unit is configured to set two access interfaces on the cloud computing center, i.e., an institutional access interface and a personal access interface; a database forming unit is configured to receive and store multidimensional medical data uploaded by the multiple medical institutions based on the distributed storage network, and form a medical database; an institutional request processing unit is configured to, when the cloud computing center receives an institutional access request of the institutional access interface, filter an initial medical data set in the medical database according to the institutional access request, wherein the institutional access request is sent by a requesting institution; an institutional data processing unit is configured to, by means of a dynamic data structure module, process the initial medical data set based on the institutional access request, generate a target medical data set, and send the target medical data set to the requesting institution; a personal request processing unit is configured to, when the cloud computing center receives a personal access request of the personal access interface, filter an associated medical data set in the medical database according to the personal access request, wherein the personal access request is sent by a requesting person; a personal report generating unit is configured to read a personal medical record of the requesting person, process the associated medical data set based on the personal access request and the personal medical record by means of a dynamic information conversion module, generate a personalized medical data analysis report, and send the personalized medical data analysis report to the requesting person. 2.The cloud-computing-based medical big data sharing platform according to claim 1, characterized in that, According to the institutional access request, filtering an initial medical data set in the medical database comprises the following steps: analyzing the institutional access request and extracting request parameters, wherein the request parameters include data type, time range and geographical range; verifying the data access authority of the requesting institution, and establishing a retrieval index in the medical database according to the request parameters; based on the retrieval index and the data access authority, filtering data meeting the conditions in the medical database, desensitizing the filtered data, and forming the initial medical data set. 3.The cloud-computing-based medical big data sharing platform according to claim 1, characterized in that, By means of a dynamic data structure module, processing the initial medical data set based on the institutional access request to generate a target medical data set comprises the following steps: deconstructing the initial medical data set into multiple data segments, extracting the structural features and semantic features of each data segment; extracting the processing requirements of the institutional access request, generating a segment combination rule based on the processing requirements, the structural features and the semantic features, recombining the multiple data segments according to the segment combination rule, and forming multiple initial medical data subsets; evaluating the relevance of the multiple initial medical data subsets to the institutional access request, iteratively adjusting the segment combination rule based on the relevance evaluation result, dynamically optimizing the recombination process, and forming multiple optimized medical data subsets; integrating the multiple optimized medical data subsets according to semantic levels, eliminating the semantic differences and format differences between the multiple optimized medical data subsets, and forming the target medical data set. 4.The cloud-computing-based medical big data sharing platform according to claim 3, characterized in that, Integrating the multiple optimized medical data subsets according to semantic levels comprises the following steps: extract semantic information of data in the plurality of optimized medical data subsets, and construct a medical semantic vector space; perform clustering analysis in the medical semantic vector space, identify semantic core areas, map the plurality of optimized medical data subsets to the most relevant semantic core areas, and realize core semantic-based data subset classification; in each semantic core area, unify the semantic representation of the data subset, adjust the data format, eliminate semantic and format differences, and then associate the data of all semantic core areas according to a preset semantic linking rule to form the target medical data set. 5.The cloud-computing-based medical big data sharing platform according to claim 1, characterized in that, According to the personal access request, the associated medical data set is screened in the medical big database, including: parsing the personal access request, extracting the personal identity, request content and request parameters; According to the personal identity, the medical records related to the requesting person are retrieved in the medical big database to obtain the personal medical archives; performing semantic analysis on the request content and combining the request parameters to construct a medical data retrieval tree; traverse the medical data retrieval tree, dynamically generate medical data screening rules based on the personal medical archives at non-leaf nodes, and use the corresponding medical data screening rules at leaf nodes to retrieve related data in the medical big database to form a preliminary personal data set; evaluate the relevance of each data in the preliminary personal data set to the personal medical archives, and remove data with a relevance lower than a preset threshold to obtain the associated medical data set. 6.The cloud-computing-based medical big data sharing platform according to claim 1, characterized in that, Through the dynamic information conversion module, the associated medical data set is processed based on the personal access request and the personal medical archives to generate a personalized medical data analysis report, including: extracting the focus points in the personal access request and the health indicators in the personal medical archives; based on the focus points and the health indicators, classifying and sorting the associated medical data set to form a plurality of theme data subsets; for each theme data subset, combining historical data in the personal medical archives to perform statistical analysis and trend prediction to generate a corresponding theme analysis result; integrate multiple theme analysis results to generate the personalized medical data analysis report, which contains personalized suggestions and risk prompts. 7.The cloud-computing-based medical big data sharing platform according to claim 6, characterized in that, For each theme data subset, combining historical data in the personal medical archives to perform statistical analysis and trend prediction to generate a corresponding theme analysis result, including: establish a personal health model based on the personal medical archives and a medical knowledge base; using the personal health model to weight the data in the theme data subset, based on the weight allocation result, performing time series analysis on the theme data subset to obtain a time series analysis result, including the change trend of key health indicators; compare the change trend of key health indicators with the expected trend in the personal health model to generate a deviation analysis result; according to the deviation analysis result, combining the medical knowledge base to generate a personalized health risk assessment; Integrating the time series analysis result, the bias analysis result and the health risk assessment forms the theme analysis result. 8.The cloud-computing-based medical big data sharing platform according to claim 1, characterized in that, Integrating multiple theme analysis results generates the personalized medical data analysis report, including: Based on the preset report template and the multiple theme analysis results, an initial report structure is generated, which contains multiple report chapters; Information extraction is performed on the multiple theme analysis results, and the extracted information is converted into text descriptions and filled into the corresponding report chapters; According to the user characteristics and preferences in the personal medical record, the filled report chapters are sorted and personalized to form the personalized medical data analysis report. According to the user characteristics and preferences in the personal medical record, the filled report chapters are sorted and personalized to form the personalized medical data analysis report.

Images