Medical information processing method and device based on cloud computing, equipment and medium

Abstract

The invention relates to the field of data processing, and particularly provides a medical information processing method and device based on cloud computing, equipment and a medium, and the method comprises the steps: constructing a medical data governance knowledge quintuple; analyzing a medical institution database table structure and field annotation through a structured knowledge ETL engine, and extracting quintuple knowledge; quintuple knowledge is extracted from PDF files such as medical industry standard documents and medical institution internal specifications through an unstructured knowledge analysis engine; the extracted quintuple is stored in a graph database in a layered mode according to entity types, and a medical data governance knowledge graph is constructed; automatically executing mapping of the manufacturer data set table and the platform data set table; mapping a manufacturer data set field and a platform data set field; mapping a manufacturer value domain and a platform value domain; when a plurality of candidate mappings exist, selecting the mapping relation with the highest confidence coefficient; feedback data in the mapping execution process are collected, and the confidence coefficient of the quintuple is adjusted. Therefore, the mapping efficiency and accuracy are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of data processing, in particular to a medical information processing method and device based on cloud computing, equipment and medium. BACKGROUND

[0002] In today's era of rapid development of digital medicine, medical data is growing explosively, covering all-round information from patient basic information, diagnosis records, test results to treatment plans. These massive and complex medical data have immeasurable value for improving medical service quality, promoting medical research, optimizing medical resource allocation, and achieving precision medicine. However, due to the wide range of medical data sources, involving different medical institutions, medical equipment manufacturers and various medical information systems, there are serious heterogeneity and inconsistency problems in data format, standard, semantics, etc., which brings great challenges to the effective integration, sharing and utilization of medical data.

[0003] With the rise of cloud computing technology, it provides strong support for large-scale data processing. Cloud computing has distributed storage capability, which can store massive medical data on multiple nodes to improve data storage efficiency and reliability; its elastic computing feature can dynamically allocate computing resources according to data processing needs to meet different requirements for computing power in medical data mapping process. At present, although some medical data governance methods and technologies have been proposed, these methods have limitations to some extent. For example, traditional data integration methods mainly focus on data extraction, transformation and loading (ETL), which converts data from different data sources into a unified format for storage and management by defining data mapping rules. However, this method usually requires manual definition of a large number of mapping rules, which is time-consuming and prone to errors, especially when dealing with complex data relationships and semantic differences, it is difficult to ensure the accuracy and completeness of the mapping. In addition, the traditional ETL method lacks tracking and management of data confidence and knowledge sources, and cannot effectively evaluate the quality of data, thereby affecting the accuracy of decisions and analysis based on these data.

[0004] As a new knowledge representation and management technology, knowledge graph can visually display entities and their relationships in the form of a graph, providing strong support for medical data integration and analysis. At present, some medical data mapping methods based on knowledge graph have emerged, but these methods mostly only focus on the simple mapping relationship between entities, lack of confidence evaluation and dynamic adjustment mechanism for mapping relationship. In practical application, due to the complexity and uncertainty of medical data, different sources of data may have conflicts and contradictions, and simple mapping methods cannot effectively handle these problems, which may lead to inaccurate and unreliable mapping results. SUMMARY

[0005] To solve the above problems, the application provides a medical information processing method, device and equipment based on cloud computing and a medium. In a first aspect, the application provides a medical information processing method based on cloud computing, which comprises the following steps: S1, constructing a medical information processing knowledge five tuple in a cloud computing environment, the format is: entity A-relation-entity B-confidence-knowledge source, wherein: the entity types of entity A and entity B include standard data set entity, data table entity, table field entity, value domain entity and quality control rule entity; the relation includes mapping relation, inheritance relation, association relation and binding relation; the confidence is used to identify the accuracy weight of automatic matching or manual confirmation; the knowledge source is used to trace the knowledge source; S2, parsing the medical institution database table structure and field annotation by deploying a structured knowledge ETL engine on the cloud computing platform, and extracting the five tuple knowledge; extracting the five tuple knowledge from the standard document PDF by deploying an unstructured knowledge parsing engine on the cloud computing platform; S3, storing the extracted five tuple according to the entity type to the graph database deployed on the cloud computing platform, and constructing a medical information processing knowledge graph; S4, based on the five tuple knowledge in the knowledge graph, using the computing resources of the cloud computing platform to automatically execute the following mappings: Mapping of manufacturer data set table and platform data set table; Mapping of manufacturer data set field and platform data set field; Mapping of manufacturer value domain and platform value domain; When there are multiple candidate mappings, the mapping relationship with the highest confidence is selected; S5, collecting feedback data in the mapping execution process, and adjusting the confidence of the related five tuple according to the feedback data.

[0006] The medical data governance knowledge five tuple is constructed, which covers various information such as entity, relation, confidence and knowledge source in a specific format, provides a comprehensive and standardized knowledge representation framework for medical data governance, helps to clearly and accurately describe the complex relationship between medical data, and is convenient for subsequent data processing and analysis. Collecting feedback data in the mapping execution process, and adjusting the confidence of the related five tuple according to the feedback data, realizes the dynamic updating and optimization of the knowledge graph, so that the mapping method can adapt to the changing medical data environment, and improve the long-term effectiveness of data governance.

[0007] As a further limitation of the application, S1 specifically comprises: S11, determining the types of entity A and entity B in the medical information processing knowledge five tuple in the cloud computing environment, the entity types include standard data set entity, data table entity, table field entity, value domain entity and quality control rule entity; S12, determine the type of relationship in the five tuple, the type of relationship includes mapping relationship, inheritance relationship, association relationship, binding relationship; S13, configure the confidence for each five tuple, set the confidence decay function, dynamically adjust the confidence of the five tuple according to the confidence decay function, and mark the five tuple as needing manual review or update if the confidence is lower than the preset threshold; S14, record the knowledge source for each five tuple, and the knowledge source is used to trace the origin of knowledge; S15, construct a medical information processing knowledge five tuple instance according to the defined entity type, relationship type, confidence and knowledge source.

[0008] The instance includes but is not limited to the following forms: (The standard data set entity, the mapping relationship, the data table entity, the confidence, and the knowledge source); (The data table entity, the inheritance relationship, the data table entity, the confidence, and the knowledge source); (The table field entity, the association relationship, the table field entity, the confidence, and the knowledge source); (The quality control rule entity, the binding relationship, the table field entity, the confidence, and the knowledge source).

[0009] Configure the confidence for each five tuple and set the confidence decay function, which can dynamically adjust the confidence of the five tuple according to the time factor. For the five tuple whose confidence is lower than the preset threshold, mark it as needing manual review or update, which ensures the reliability and timeliness of the knowledge in the knowledge graph, and timely discovers and processes the knowledge that may exist errors or be outdated. Record the detailed knowledge source for each five tuple, including various possible origins, which is convenient for tracing the original basis of knowledge when needed, helps to review, verify and update the knowledge, and improves the credibility and maintainability of data governance.

[0010] According to the defined entity type, relationship type, confidence and knowledge source, construct a medical data governance knowledge five tuple instance, which ensures that each five tuple conforms to the unified standard and specification, and provides a standardized data basis for subsequent knowledge extraction, storage and mapping operation.

[0011] As a further limitation of the application, in S13, the confidence decay function is:

[0012] In the formula, is the attenuation coefficient, , is the half-life parameter, which represents the time required for the confidence to decay to half of the initial value, is a time variable, which represents the time since the confidence is set or adjusted.

[0013] Through the parameters such as attenuation coefficient, half-life parameter and time variable, the change trend of the confidence with time can be accurately simulated. This model can adjust the parameters according to the actual situation to adapt to the attenuation characteristics of different types of knowledge, and provides an effective mathematical tool for dynamically managing the confidence of the five-tuple.

[0014] Based on the confidence attenuation function, the confidence of the five-tuple is dynamically adjusted, which can timely find those knowledge that may become inaccurate over time, and through marking for manual review or updating, prompt relevant personnel to re-evaluate and correct the knowledge, thereby improving the accuracy and reliability of the knowledge in the knowledge graph.

[0015] As a further limitation of the application, S2 specifically comprises: By deploying a structured knowledge ETL engine on a cloud computing platform, the DDL statement and data dictionary of the medical institution database are parsed, the three-level structure of table-field-value domain is extracted, and the structured five-tuple is generated; By deploying an integrated OCR and large language model on a cloud computing platform, an unstructured parsing engine is used to identify data element definitions and business rules from standard PDFs, and unstructured five-tuples are generated.

[0016] By parsing the DDL statement and data dictionary of the medical institution database through the structured knowledge ETL engine, the three-level structure of table-field-value domain is extracted and the structured five-tuple is generated, which can quickly and accurately obtain key information from structured data sources, improve the efficiency and accuracy of knowledge extraction, and provide rich structured knowledge for building a knowledge graph.

[0017] The integrated OCR and large language model unstructured parsing engine can identify data element definitions and business rules from standard PDFs and generate unstructured five-tuples. This technology combines the text recognition capabilities of OCR and the natural language processing capabilities of large language models, effectively solving the problem of knowledge extraction in unstructured documents, and further enriching the data sources of the knowledge graph.

[0018] As a further limitation of the application, S3 specifically comprises: S31, classifying the extracted five-tuples, and according to the entity types of entity A and entity B, the five-tuples are classified into five-tuples containing standard data set entities, five-tuples containing data table entities, five-tuples containing table field entities, five-tuples containing value domain entities, and five-tuples containing quality control rule entities; S32, creating a plurality of hierarchical storage areas in the graph database, each storage area corresponding to an entity type; storing the classified five-tuples into the corresponding hierarchical storage area; S33, establish the relationship between the five tuples in the graph database, specifically including: According to the relationship type in the five tuple, create a corresponding edge in the graph database; add confidence and knowledge source attributes to each edge to ensure that each relationship in the knowledge graph has a clear confidence and source; S34, verify the constructed knowledge graph to ensure that the relationship between the five tuples is logically correct, specifically including: Check if the relationship between the five tuples conforms to the predefined logical rules; Check if there are isolated entities or relationships; S35, optimize the knowledge graph according to the verification result, specifically including: Correct the relationship that does not conform to the logic; Supplement the missing five tuples or relationships; Adjust the confidence parameters of the five tuples whose confidence is below the set threshold.

[0019] Classify the extracted five tuples and create multiple hierarchical storage areas in the graph database, store the five tuples of different entity types into the corresponding areas, realize the ordered classification and storage of knowledge. This hierarchical storage method facilitates knowledge management and query, improves data access efficiency, and also provides convenience for subsequent knowledge graph construction and maintenance.

[0020] In the graph database, create a corresponding edge according to the relationship type in the five tuple, and add confidence and knowledge source attributes to each edge to ensure that each relationship in the knowledge graph has a clear confidence and source. This not only helps to understand the relationship between entities, but also can reasonably select according to the confidence in the subsequent mapping operation, improve the accuracy of mapping.

[0021] Verify the constructed knowledge graph to check if the relationship between the five tuples conforms to the predefined logical rules and if there are isolated entities or relationships, which can timely find the problems and errors in the knowledge graph, ensure the logical correctness and integrity of the knowledge graph, and provide a reliable foundation for subsequent data governance operations.

[0022] Optimize the knowledge graph according to the verification result, including correcting the relationship that does not conform to the logic, supplementing the missing five tuples or relationships, and adjusting the confidence parameters of the five tuples whose confidence is below the set threshold, which can continuously improve the quality and accuracy of the knowledge graph, and make it better adapt to the needs of medical data governance.

[0023] As a further limitation of the present application, the classified five tuples are stored in the corresponding hierarchical storage area, specifically including: Store the five tuples of the standard dataset class in the standard dataset layer; store the five-tuples of the data table class to the data table layer; store the five-tuples of the table field class to the table field layer; store the five-tuples of the value domain class to the value domain layer; store the five-tuples of the quality control rule class to the quality control rule layer.

[0024] The five-tuples of different categories are respectively stored to the standard data set layer, the data table layer, the table field layer, the value domain layer and the quality control rule layer. This clear hierarchical storage structure makes the knowledge organization in the graph database more reasonable, facilitates quick retrieval and management according to different entity types, and improves the efficiency and accuracy of data processing. Different levels of storage areas correspond to different entity types. When performing data governance operations, specific levels of data can be processed and analyzed, such as table structure optimization of data in the data table layer, value domain specification inspection of data in the value domain layer, etc., improving the pertinence and effectiveness of data governance.

[0025] As a further limitation of the application, S4 specifically includes: S41, extract all related five-tuple knowledge from the knowledge graph, including five-tuples of the manufacturer data set table, the platform data set table, the manufacturer data set field, the platform data set field, the manufacturer value domain and the platform value domain; S42, map the manufacturer data set table and the platform data set table, specifically including: calculate the similarity between the manufacturer data set table and the platform data set table, the similarity calculation is based on table name, table annotation and table structure; select the best matching table mapping relationship according to the similarity and the confidence; S43, map the manufacturer data set field and the platform data set field, specifically including: calculate the similarity between the manufacturer data set field and the platform data set field, the similarity calculation is based on field name, field annotation, field type and field value domain; select the best matching field mapping relationship according to the similarity and the confidence; S44, map the manufacturer value domain and the platform value domain, specifically including: calculate the similarity between the manufacturer value domain and the platform value domain, the similarity calculation is based on the value range and the enumeration value of the value domain; select the best matching value domain mapping relationship according to the similarity and the confidence; S45, for the case where there are multiple candidate mappings, the following steps are performed: compare the confidence of the candidate mappings and select the mapping relationship with the highest confidence; if the confidence is the same, compare the reliability of the knowledge sources and preferentially select the mapping relationship with more reliable knowledge sources; If the knowledge sources are also the same, the final mapping relationship is determined by manual review.

[0026] All relevant five-tuple knowledge is extracted from the knowledge graph, including data information of each level of the manufacturer and the platform, to provide comprehensive knowledge support for the mapping operation, ensuring that the data mapping between the manufacturer and the platform can be accurately and completely completed. In the mapping process of tables, fields and value ranges, the similarity is calculated based on different characteristics respectively, and the best matching mapping relationship is selected according to the similarity and the confidence, and this method of comprehensively considering multiple factors can improve the accuracy and reliability of the mapping and reduce the occurrence of incorrect mapping. For the case where there are multiple candidate mappings, the final mapping relationship is determined by comparing the confidence, the reliability of the knowledge source and the manual review, and this multi-level processing strategy can make full use of the existing knowledge information, ensure that the most suitable mapping relationship can be selected in complex situations, and improve the robustness and adaptability of the mapping method.

[0027] As a further limitation of the application, S5 specifically includes: S51, in the mapping execution process, collecting feedback data of users on the mapping result, the feedback data including at least one of the following forms: user confirmation mark of the mapping relationship; user correction suggestion of the mapping relationship; user confidence evaluation of the mapping relationship; verification result of the mapping relationship by an automatic verification tool; S52, classifying the collected feedback data, and classifying the feedback data into the following categories according to the feedback type: confirmation feedback: the user confirms that the mapping relationship is correct; correction feedback: the user proposes a correction suggestion for the mapping relationship; evaluation feedback: the user evaluates the confidence of the mapping relationship; verification feedback: the verification result of the mapping relationship by an automatic verification tool; S53, according to the classified feedback data, adjusting the confidence of the related five-tuple according to the following rules: for the confirmation feedback, increase the confidence of the related five-tuple; for the correction feedback, reduce the confidence of the related five-tuple, and update the five-tuple according to the correction suggestion; for the evaluation feedback, adjust the confidence of the related five-tuple according to the confidence evaluation value given by the user; for the verification feedback, adjust the confidence of the related five-tuple according to the verification result, and increase the confidence if the verification is passed, and reduce the confidence if the verification fails; S54. Based on the above adjustment rules, the confidence level of the relevant quintuples is specifically adjusted. The adjusted confidence level value ranges from 0 to 1, where 1 represents complete confirmation and 0 represents complete uncertainty. For the quintuples with adjusted confidence levels, the corresponding records in the knowledge graph are updated. S55. Record each feedback data and its corresponding confidence adjustment results to form a feedback history record; periodically analyze the feedback history record to evaluate the overall quality and reliability of the knowledge graph; based on the analysis results, optimize the construction and updating strategy of the knowledge graph to further improve the accuracy and reliability of the mapping.

[0028] Secondly, the present invention provides a cloud computing-based medical information processing device, comprising: A knowledge quintuple construction module, deployed on a cloud computing platform, is used to construct knowledge quintuples for medical information processing. The format of the quintuple is: Entity A - Relationship - Entity B - Confidence - Knowledge Source. Entity A and Entity B are independently selected from the following entity types: standard dataset entity, data table entity, table field entity, value range entity, and quality control rule entity. Relationships include mapping relationships, inheritance relationships, association relationships, and binding relationships. Confidence is used to identify the accuracy weight of automatic matching or manual confirmation. Knowledge source is used to trace the origin of knowledge. The knowledge extraction module, deployed on a cloud computing platform, is used to parse the table structure and field annotations of medical institution databases using a structured knowledge ETL engine to extract five-tuple knowledge; and to extract five-tuple knowledge from PDF files such as medical industry standard documents and internal regulations of medical institutions using an unstructured knowledge parsing engine. The knowledge storage module, deployed on a cloud computing platform, is used to store the extracted quintuples hierarchically in a graph database according to entity type, thereby constructing a medical information processing knowledge graph. The mapping execution module, deployed on a cloud computing platform, is used to automatically execute the following mappings based on the quintuple knowledge in the knowledge graph: Mapping between manufacturer dataset tables and platform dataset tables; Mapping between manufacturer dataset fields and platform dataset fields; Mapping between manufacturer value domains and platform value domains; When multiple candidate mappings exist, the mapping with the highest confidence level is selected; The feedback processing module, deployed on a cloud computing platform, is used to collect feedback data during the mapping execution process and adjust the confidence of relevant quintuples based on the feedback data.

[0029] As a further limitation of the present invention, the knowledge quintuple construction module includes: The entity type determination unit is used to determine the types of entity A and entity B in the medical data governance knowledge quintuple. The entity types are independently selected from the following list: standard dataset entity, data table entity, table field entity, value range entity, and quality control rule entity. The relation type determination unit is used to determine the type of relation in the quintuple. The types of relations include mapping relations, inheritance relations, association relations, and binding relations. The confidence level allocation unit is used to configure the confidence level for each quintuple, set the confidence level decay function, and dynamically adjust the confidence level of the quintuple according to the confidence level decay function. For quintuples with a confidence level lower than a preset threshold, they are marked as requiring manual review or updating. The knowledge source recording unit is used to record the knowledge source for each quintuple. The knowledge source is used to trace the origin of the knowledge, including but not limited to standard documents, platform standard documents, vendor standard documents, business annotation data, dataset quality control rules, and medical business specifications. The quintuple instance building unit is used to construct medical data governance knowledge quintuple instances based on the defined entity type, relation type, confidence level, and knowledge source.

[0030] Thirdly, the present invention also provides an electronic device, the electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; the memory storing computer program instructions executable by the at least one processor, the computer program instructions being executed by the at least one processor to enable the at least one processor to perform the cloud-based medical information processing method as described in the first aspect.

[0031] Fourthly, the present invention also provides a non-transitory computer-readable storage medium that stores computer instructions that cause the computer to execute the cloud-based medical information processing method as described in the first aspect.

[0032] As can be seen from the above technical solutions, this application has the following advantages: By using a structured knowledge ETL engine to parse the table structure and field annotations of the medical institution database, and an unstructured knowledge parsing engine to extract quintuple knowledge from standard document PDFs, it achieves the acquisition of medical data governance knowledge from multiple different sources, enriching the data foundation of the knowledge graph and improving the completeness and accuracy of the knowledge. The extracted quintuples are stored hierarchically in a graph database according to entity type, constructing a medical data governance knowledge graph. This storage method facilitates the classification, management, and rapid retrieval of knowledge. Simultaneously, the graph structure of the graph database can intuitively display the relationships between entities, providing efficient support for subsequent mapping operations.

[0033] Based on the five-tuple knowledge in the knowledge graph, the system automatically performs the mapping of dataset tables, dataset fields, and value ranges between manufacturers and the platform, improving the efficiency and accuracy of the mapping. When multiple candidate mappings exist, the mapping relationship with the highest confidence is selected, further ensuring the reliability of the mapping results and reducing the workload of manual intervention. Attached Figure Description

[0034] To more clearly illustrate the technical solution of this application, the accompanying drawings used in the description will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart illustrating the method provided in an embodiment of the present invention.

[0036] Figure 2 A block diagram of the apparatus provided in an embodiment of the present invention. Detailed Implementation

[0037] To make the purpose, features, and advantages of this application more apparent and understandable, specific embodiments and accompanying drawings will be used to clearly and completely describe the technical solution protected by this application. Obviously, the embodiments described below are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this application and in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0039] like Figure 1 As shown, this embodiment of the invention provides a cloud computing-based medical information processing method, including the following steps: S1. Construct a medical information processing knowledge quintuple in a cloud computing environment, in the format: Entity A - Relationship - Entity B - Confidence - Knowledge Source, wherein: the entity types of Entity A and Entity B include standard dataset entities, data table entities, table field entities, value domain entities, and quality control rule entities; the relationship includes mapping relationships, inheritance relationships, association relationships, and binding relationships; the confidence level is used to identify the accuracy weight of automatic matching or manual confirmation; the knowledge source includes, but is not limited to, medical industry standard documents, medical institution database dictionaries, dataset descriptions provided by vendors, and manually annotated records, etc., used to trace the original source of knowledge; S1 specifically includes: S11. In a cloud computing environment, determine the types of entity A and entity B in the medical information processing knowledge quintuple, wherein the entity types include standard dataset entity, data table entity, table field entity, value domain entity, and quality control rule entity; Examples: Standard dataset entity (such as medical standard dataset), data table entity (such as patient information table in hospital information system), table field entity (such as patient name field), value range entity (such as the value range "male / female" of gender field), quality control rule entity (such as data integrity verification rules).

[0040] S12. Determine the type of relation in the quintuple. The types of relations include mapping, inheritance, association, and binding. Examples: mapping (such as the mapping between fields in the standard dataset and fields in the hospital data table), inheritance (such as the child table inheriting fields from the parent table), association (such as the association between the patient table and the diagnosis table), and binding (such as the binding of quality control rules to specific fields).

[0041] S13. Configure confidence level for each quintuple component, set confidence level decay function, and dynamically adjust the confidence level of the quintuple according to the confidence level decay function. For quintuples with confidence level lower than a preset threshold (e.g., 0.5), mark them as needing manual review or updating. Confidence level is used to identify the accuracy weight of automatic matching or manual confirmation, and ranges from 0 to 1, where 1 represents complete confirmation and 0 represents no confirmation. The confidence decay function is:

[0042] In the formula, The attenuation coefficient is... , The half-life parameter represents the time required for the confidence level to decay to half of its initial value. This is a time variable, representing the time since the confidence level was set or adjusted.

[0043] S14. Record the knowledge source for each quintuple. The knowledge source is used to trace the origin of the knowledge, including but not limited to standard documents, platform standard documents, vendor standard documents, business annotation data, dataset quality control rules, and medical business specifications. For example, the knowledge source may be a medical standard document, a hospital information system data dictionary, or a data standard document provided by a vendor.

[0044] S15. Based on the defined entity type, relation type, confidence level, and knowledge source, construct a medical data governance knowledge quintuple instance.

[0045] The examples include, but are not limited to, the following forms: (Standard dataset entities, mapping relationships, data table entities, confidence level 0.9, medical standard documents); (Data table entity, inheritance relationship, data table entity, confidence level 0.8, hospital information system data dictionary); (Table field entity, relationship, table field entity, confidence level 0.7, business-labeled data); (Quality control rule entity, binding relationship, table field entity, confidence level 0.6, dataset quality control rule).

[0046] S2. Using a structured knowledge ETL engine deployed on a cloud computing platform, the table structure and field annotations of the medical institution's database are parsed to extract five-tuple knowledge; a non-structured knowledge parsing engine deployed on a cloud computing platform is used to extract five-tuple knowledge from standard document PDFs; S2 specifically includes: By using a structured knowledge ETL engine deployed on a cloud computing platform to parse the DDL statements and data dictionary of the medical institution's database, the three-level structure of table-field-value domain is extracted to generate structured quintuples. Configure database connection parameters (database type (e.g., MySQL, Oracle, SQL Server), database address, port, username, password, database name, and table name), and establish a connection using a database connection library (e.g., pymysql, psycopg2). Check the connection success and record the connection status. Use SQL queries to extract the structure information of the database tables, including table names, field names, field types, and field comments. The data dictionary typically contains detailed descriptions of the tables and the business meanings of the fields. If the data dictionary is stored in file format (e.g., Excel, CSV), use a suitable parsing tool (e.g., pandas) to read the data dictionary. Value range information may be stored in specific tables or explicitly specified in field comments.

[0047] Based on the extracted table structure, field information, and value range information, construct a structured quintuple.

[0048] The format of a quintuple is: (Entity A, Relation, Entity B, Confidence, Source of Knowledge).

[0049] In this embodiment of the invention, by connecting to the hospital information system (HIS) database, the following is obtained: (Data table entity, containment relationship, table field entity, 0.9, HIS database) (Table field entities, value range relationships, value range entities, 0.8, HIS database) Traverse the table structure information and generate a 5-tuple for each table and field. Store the generated 5-tuples in a graph database or other storage system.

[0050] By deploying an unstructured parsing engine that integrates OCR and a large language model on a cloud computing platform, data element definitions and business rules are identified from standard PDFs, generating unstructured quintuples.

[0051] For example, extracting data meta definitions from a medical standards PDF document generates the following quintuple: (Standard dataset entities, defined relations, data element entities, 0.9, medical standard documents) (Data element entity, value domain relationship, value domain entity, 0.8, medical standard document) S3. The extracted quintuples are stored hierarchically according to entity type in a graph database deployed on a cloud computing platform to construct a medical information processing knowledge graph; S3 specifically includes: S31. Classify the extracted quintuples according to the entity types of entity A and entity B. Divide the quintuples into quintuples containing standard dataset entities, quintuples containing data table entities, quintuples containing table field entities, quintuples containing value range entities, and quintuples containing quality control rule entities. S32. Create multiple hierarchical storage areas in the graph database, with each storage area corresponding to a type of entity; store the categorized quintuples into the corresponding hierarchical storage areas. S33. Establish relationships between quintuples in a graph database, specifically including: Based on the relation type in the quintuple, create corresponding edges in the graph database; add confidence and knowledge source attributes to each edge to ensure that each relation in the knowledge graph has a clear confidence and source; S34. Verify the constructed knowledge graph to ensure the logical correctness of the relationships between the quintuples, specifically including: Check whether the relationships between the quintuples conform to predefined logical rules; Check for the existence of isolated entities or relationships; S35. Optimize the knowledge graph based on the verification results, specifically including: Correct illogical relationships; Supplement missing quintuples or relations; Adjust the confidence parameter of quintuples whose confidence level is lower than the set threshold.

[0052] The categorized quintuples are stored in the corresponding hierarchical storage areas, specifically including: Store the quintuples of the standard dataset class into the standard dataset layer; Store the quintuple of the data table class in the data table layer; Store the quintuples of the table field class in the table field layer; Store the quintuple of the range class in the range layer; Store the quintuple of the quality control rule class in the quality control rule layer.

[0053] S4. Based on the quintuple knowledge in the knowledge graph, automatically perform the following mapping using the computing resources of the cloud computing platform: Mapping between manufacturer dataset tables and platform dataset tables; Mapping between manufacturer dataset fields and platform dataset fields; Mapping between manufacturer value domains and platform value domains; When there are multiple candidate mappings, select the mapping with the highest confidence level; S4 specifically includes: S41. Extract all relevant quintuple knowledge from the knowledge graph, including quintuples of manufacturer dataset table, platform dataset table, manufacturer dataset field, platform dataset field, manufacturer value domain, and platform value domain; S42. Map the manufacturer's dataset table to the platform's dataset table, specifically including: Calculate the similarity between the manufacturer's dataset table and the platform's dataset table. The similarity calculation is based on the table name, table comments, and table structure. The best matching table mapping relationship is selected based on similarity and confidence. S43. Map the manufacturer's dataset fields to the platform's dataset fields, specifically including: Calculate the similarity between fields in the manufacturer's dataset and fields in the platform dataset. The similarity calculation is based on field name, field annotation, field type, and field value range. Select the best matching field mapping relationship based on similarity and confidence; S44. Map the manufacturer's value domain to the platform's value domain, specifically including: Calculate the similarity between the manufacturer's value range and the platform's value range. The similarity calculation is based on the value range and enumerated values ​​of the value range. The best matching value range mapping relationship is selected based on similarity and confidence. S45. For cases where there are multiple candidate mappings, perform the following steps: Compare the confidence levels of the candidate mappings and select the mapping relationship with the highest confidence level; If the confidence levels are the same, compare the reliability of the knowledge sources and prioritize the mapping relationship with the more reliable knowledge source; If the knowledge sources are also the same, the final mapping relationship is determined through manual review.

[0054] S5. Collect feedback data during the mapping process and adjust the confidence levels of relevant quintuples based on the feedback data. S5 specifically includes: S51. During the mapping process, user feedback data on the mapping results is collected, and the feedback data includes at least one of the following forms: User's confirmation marker regarding the mapping relationship; User suggestions for correcting the mapping relationship; User confidence assessment of the mapping relationship; The results of the automatic verification tool on the mapping relationship; S52. Classify the collected feedback data according to the feedback type, and divide the feedback data into the following categories: Confirmation feedback: The user confirms that the mapping relationship is correct; Corrective feedback: User suggestions for correcting the mapping relationship; Evaluation-type feedback: Users evaluate the confidence level of the mapping relationship; Validation feedback: The results of the automatic validation tool's validation of the mapping relationship; S53. Based on the categorized feedback data, adjust the confidence level of the relevant quintuples according to the following rules: For confirmatory feedback, increase the confidence level of the relevant quintuple; For corrective feedback, reduce the confidence level of the relevant quintuple and update the quintuple based on the corrective suggestions; For assessment-type feedback, adjust the confidence level of the relevant quintuples based on the confidence level assessment value provided by the user; For validation feedback, adjust the confidence level of the relevant quintuples based on the validation results; increase the confidence level if the validation passes and decrease the confidence level if the validation fails. S54. Based on the above adjustment rules, the confidence level of the relevant quintuples is specifically adjusted. The adjusted confidence level value ranges from 0 to 1, where 1 represents complete confirmation and 0 represents complete uncertainty. For the quintuples with adjusted confidence levels, the corresponding records in the knowledge graph are updated. S55. Record each feedback data and its corresponding confidence adjustment results to form a feedback history record; periodically analyze the feedback history record to evaluate the overall quality and reliability of the knowledge graph; based on the analysis results, optimize the construction and updating strategy of the knowledge graph to further improve the accuracy and reliability of the mapping.

[0055] like Figure 2 As shown, this embodiment of the invention provides a cloud computing-based medical information processing device, comprising: A knowledge quintuple construction module, deployed on a cloud computing platform, is used to construct knowledge quintuples for medical information processing. The format of the quintuple is: Entity A - Relationship - Entity B - Confidence - Knowledge Source. Entity A and Entity B are independently selected from the following entity types: standard dataset entity, data table entity, table field entity, value range entity, and quality control rule entity. Relationships include mapping relationships, inheritance relationships, association relationships, and binding relationships. Confidence is used to identify the accuracy weight of automatic matching or manual confirmation. Knowledge source is used to trace the origin of knowledge. The knowledge extraction module, deployed on a cloud computing platform, is used to parse the table structure and field annotations of medical institution databases using a structured knowledge ETL engine to extract five-tuple knowledge; and to extract five-tuple knowledge from PDF files such as medical industry standard documents and internal regulations of medical institutions using an unstructured knowledge parsing engine. The knowledge storage module, deployed on a cloud computing platform, is used to store the extracted quintuples hierarchically in a graph database according to entity type, thereby constructing a medical information processing knowledge graph. The mapping execution module, deployed on a cloud computing platform, is used to automatically execute the following mappings based on the quintuple knowledge in the knowledge graph: Mapping between manufacturer dataset tables and platform dataset tables; Mapping between manufacturer dataset fields and platform dataset fields; Mapping between manufacturer value domains and platform value domains; When multiple candidate mappings exist, the mapping with the highest confidence level is selected; The feedback processing module, deployed on a cloud computing platform, is used to collect feedback data during the mapping execution process and adjust the confidence of relevant quintuples based on the feedback data.

[0056] In some embodiments, the knowledge quintuple construction module includes: The entity type determination unit is used to determine the types of entity A and entity B in the medical data governance knowledge quintuple. The entity types are independently selected from the following list: standard dataset entity, data table entity, table field entity, value range entity, and quality control rule entity. The relation type determination unit is used to determine the type of relation in the quintuple. The types of relations include mapping relations, inheritance relations, association relations, and binding relations. The confidence level allocation unit is used to configure the confidence level for each quintuple, set the confidence level decay function, and dynamically adjust the confidence level of the quintuple according to the confidence level decay function. For quintuples with a confidence level lower than a preset threshold, they are marked as requiring manual review or updating. The knowledge source recording unit is used to record the knowledge source for each quintuple. The knowledge source is used to trace the origin of the knowledge, including but not limited to standard documents, platform standard documents, vendor standard documents, business annotation data, dataset quality control rules, and medical business specifications. The quintuple instance building unit is used to construct medical data governance knowledge quintuple instances based on the defined entity type, relation type, confidence level, and knowledge source.

[0057] In some embodiments, the knowledge extraction module includes: The structured knowledge extraction unit is used to parse the DDL statements and data dictionary of the medical institution database through the structured knowledge ETL engine, extract the three-level structure of table-field-value domain, and generate structured quintuples. The unstructured knowledge extraction unit is used to identify data element definitions and business rules from standard PDFs by integrating an unstructured parsing engine with OCR and a large language model, and generate unstructured quintuples.

[0058] In some embodiments, the knowledge storage module includes: The quintuple classification unit is used to classify the extracted quintuples. Based on the entity types of entity A and entity B, the quintuples are divided into quintuples containing standard dataset entities, quintuples containing data table entities, quintuples containing table field entities, quintuples containing value range entities, and quintuples containing quality control rule entities. Hierarchical storage units are used to create multiple hierarchical storage areas in a graph database, with each storage area corresponding to a type of entity; the categorized quintuples are stored in the corresponding hierarchical storage areas. The relation-building unit is used to establish relationships between quintuples in a graph database, specifically including: Based on the relation type in the quintuple, create the corresponding edge in the graph database; Add confidence and knowledge source attributes to each edge to ensure that every relationship in the knowledge graph has a clear confidence level and source; The knowledge graph verification unit is used to verify the constructed knowledge graph, ensuring the logical correctness of the relationships between the quintuples. Specifically, it includes: Check whether the relationships between the quintuples conform to predefined logical rules; Check for the existence of isolated entities or relationships; The knowledge graph optimization unit is used to optimize the knowledge graph based on the validation results, specifically including: Correct illogical relationships; Supplement missing quintuples or relations; Adjust the confidence parameter of quintuples whose confidence level is lower than the set threshold.

[0059] In some embodiments, the hierarchical storage unit specifically includes: The standard dataset layer storage unit is used to store the five-tuple of the standard dataset class in the standard dataset layer; The data table layer storage unit is used to store the five-tuple of the data table class in the data table layer; The table field layer storage unit is used to store the five-tuple of the table field class in the table field layer; The value range layer storage unit is used to store the quintuple of the value range class in the value range layer; The quality control rule layer storage unit is used to store the five-tuple of the quality control rule class into the quality control rule layer.

[0060] In some embodiments, the mapping execution module includes: The knowledge extraction unit is used to extract all relevant quintuple knowledge from the knowledge graph, including quintuples of manufacturer dataset table, platform dataset table, manufacturer dataset field, platform dataset field, manufacturer value domain, and platform value domain; The table mapping unit is used to map the manufacturer's dataset table to the platform's dataset table, specifically including: Calculate the similarity between the manufacturer's dataset table and the platform's dataset table. The similarity calculation is based on the table name, table comments, and table structure. The best matching table mapping relationship is selected based on similarity and confidence. The field mapping unit is used to map fields in the manufacturer's dataset to fields in the platform's dataset. Specifically, it includes: Calculate the similarity between fields in the manufacturer's dataset and fields in the platform dataset. The similarity calculation is based on field name, field annotation, field type, and field value range. Select the best matching field mapping relationship based on similarity and confidence; The value range mapping unit is used to map the manufacturer's value range to the platform's value range, and specifically includes: Calculate the similarity between the manufacturer's value range and the platform's value range. The similarity calculation is based on the value range and enumerated values ​​of the value range. The best matching value range mapping relationship is selected based on similarity and confidence. The conflict resolution unit performs the following steps when there are multiple candidate mappings: Compare the confidence levels of the candidate mappings and select the mapping relationship with the highest confidence level; If the confidence levels are the same, compare the reliability of the knowledge sources and prioritize the mapping relationship with the more reliable knowledge source; If the knowledge sources are also the same, the final mapping relationship is determined through manual review.

[0061] This invention also provides an electronic device, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, communication interface, and memory communicate with each other via the communication bus. The communication bus can be used for information transmission between the electronic device and sensors. The processor can call logical instructions in the memory to execute the following method: S1, constructing a medical information processing knowledge quintuple in a cloud computing environment, in the format: Entity A-Relationship-Entity B-Confidence-Knowledge Source, wherein: the entity types of Entity A and Entity B include standard dataset entities, data table entities, table field entities, value domain entities, and quality control rule entities; the relationships include mapping relationships, inheritance relationships, association relationships, and binding relationships; the confidence level is used to identify the accuracy weight of automatic matching or manual confirmation; the knowledge source is used to trace the source of knowledge; S2, parsing the medical institution database table structure and field annotations using a structured knowledge ETL engine deployed on a cloud computing platform to extract the quintuple knowledge. The process involves: S3, extracting quintuple knowledge from standard PDF documents using an unstructured knowledge parsing engine deployed on a cloud computing platform; S4, storing the extracted quintuples hierarchically by entity type in a graph database deployed on the cloud computing platform to construct a medical information processing knowledge graph; S5, automatically performing the following mappings based on the quintuple knowledge in the knowledge graph, utilizing the computing resources of the cloud computing platform: mapping between the manufacturer's dataset table and the platform's dataset table; mapping between the manufacturer's dataset fields and the platform's dataset fields; mapping between the manufacturer's value domain and the platform's value domain; and selecting the mapping relationship with the highest confidence when multiple candidate mappings exist; and S6, collecting feedback data during the mapping process and adjusting the confidence of the relevant quintuples based on the feedback data.

[0062] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and sold or used as independent products, and can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0063] This invention provides a non-transitory computer-readable storage medium storing computer instructions that cause a computer to execute the methods provided in the above-described method embodiments. For example, the instructions include: S1, constructing a medical information processing knowledge quintuple in a cloud computing environment, in the format: Entity A-Relationship-Entity B-Confidence-Knowledge Source, wherein: the entity types of Entity A and Entity B include standard dataset entities, data table entities, table field entities, value domain entities, and quality control rule entities; the relationships include mapping relationships, inheritance relationships, association relationships, and binding relationships; the confidence level is used to identify the accuracy weight of automatic matching or manual confirmation; and the knowledge source is used to trace the origin of knowledge; S2, using a structured knowledge ETL engine deployed on a cloud computing platform. The process involves: S1) parsing the table structure and field annotations of the medical institution database to extract quintuple knowledge; S2) extracting quintuple knowledge from standard PDF documents using an unstructured knowledge parsing engine deployed on a cloud computing platform; S3) storing the extracted quintuples hierarchically according to entity type in a graph database deployed on the cloud computing platform to construct a medical information processing knowledge graph; S4) automatically performing the following mappings based on the quintuple knowledge in the knowledge graph, utilizing the computing resources of the cloud computing platform: mapping between the manufacturer's dataset table and the platform's dataset table; mapping between the manufacturer's dataset fields and the platform's dataset fields; mapping between the manufacturer's value domain and the platform's value domain; and selecting the mapping relationship with the highest confidence when multiple candidate mappings exist; and S5) collecting feedback data during the mapping process and adjusting the confidence of relevant quintuples based on the feedback data.

[0064] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cloud computing-based medical information processing method, characterized by, Comprise the following steps: S1, build medical information processing knowledge five tuple in cloud computing environment, format is: entity A-relation-entity B-confidence-knowledge source, wherein: the entity type of entity A and entity B includes standard dataset entity, data table entity, table field entity, value domain entity, quality control rule entity; relation includes mapping relation, inheritance relation, association relation, binding relation; confidence is used for identifying the accuracy weight of automatic matching or manual confirmation; knowledge source is used for tracing the knowledge source; S2, parse the table structure and field annotation of the medical institution database through the structured knowledge ETL engine deployed on the cloud computing platform, and extract the five tuple knowledge; extract the five tuple knowledge from the standard document PDF through the unstructured knowledge parsing engine deployed on the cloud computing platform; S3, store the extracted five tuple to the graph database deployed on the cloud computing platform according to the entity type, and build medical information processing knowledge graph; S4, based on the five tuple knowledge in the knowledge graph, use the computing resources of the cloud computing platform to automatically execute the following mapping: Mapping of manufacturer dataset table and platform dataset table; Mapping of manufacturer dataset field and platform dataset field; Mapping of manufacturer value domain and platform value domain; When there are multiple candidate mappings, select the mapping relationship with the highest confidence; S5, collect feedback data in the mapping execution process, and adjust the confidence of the related five tuple according to the feedback data.

2. The cloud-computing-based medical information processing method of claim 1, wherein, S1 specifically includes: S11, determine the type of entity A and entity B in the medical information processing knowledge five tuple in the cloud computing environment, the entity type includes standard dataset entity, data table entity, table field entity, value domain entity, quality control rule entity; S12, determine the type of relation in the five tuple, the type of relation includes mapping relation, inheritance relation, association relation, binding relation; S13, configure the confidence for each five tuple, set the confidence decay function, and dynamically adjust the confidence of the five tuple according to the confidence decay function, mark the five tuple with confidence lower than the preset threshold as needing manual review or update; S14, record the knowledge source for each five tuple, the knowledge source is used for tracing the source of knowledge; S15, build medical information processing knowledge five tuple instance according to the defined entity type, relation type, confidence and knowledge source. 3.The cloud-computing-based medical information processing method of claim 2, wherein, In S13, the confidence decay function is: wherein is a decay coefficient, , is a half-life parameter, representing the time required for the confidence to decay to half of the initial value, is a time variable, representing the time since the confidence was set or adjusted. 4.The cloud-computing-based medical information processing method of claim 3, wherein, S2 specifically includes: Parse the DDL statement and data dictionary of the medical institution database through the structured knowledge ETL engine deployed on the cloud computing platform, extract the three-level structure of table-field-value domain, and generate structured five tuple; Through the unstructured parsing engine integrated with OCR and large language model deployed on the cloud computing platform, identify data element definition and business rules from standard PDF, and generate unstructured five tuple. 5.The cloud-computing-based medical information processing method of claim 4, wherein, S3 specifically includes: S31, classify the extracted five tuple, and divide the five tuple into five tuple containing standard dataset entity, five tuple containing data table entity, five tuple containing table field entity, five tuple containing value domain entity, and five tuple containing quality control rule entity according to the entity type of entity A and entity B; S32, create multiple hierarchical storage areas in the graph database deployed on the cloud computing platform, each corresponding to an entity type; store the classified five-tuples into the corresponding hierarchical storage areas; S33, establish relationships between five-tuples in the graph database, specifically including: According to the relationship type in the five-tuple, create the corresponding edge in the graph database; Add confidence and knowledge source attributes to each edge to ensure that each relationship in the knowledge graph has a clear confidence and source; S34: Verify the constructed knowledge graph to ensure that the relationships between five-tuples are logically correct, specifically including: Check if the relationships between five-tuples conform to the predefined logical rules; Check if there are isolated entities or relationships; S35, optimize the knowledge graph according to the verification results, specifically including: Correct the relationships that do not conform to the logic; Supplement the missing five-tuples or relationships; Adjust the confidence parameters of five-tuples with confidence below the set threshold. 6.The cloud-computing-based medical information processing method of claim 5, wherein, Store the classified five-tuples into the corresponding hierarchical storage areas, specifically including: Store the five-tuples of the standard dataset class into the standard dataset layer; Store the five-tuples of the data table class into the data table layer; Store the five-tuples of the table field class into the table field layer; Store the five-tuples of the value domain class into the value domain layer; Store the five-tuples of the quality control rule class into the quality control rule layer. 7.The cloud-computing-based medical information processing method of claim 6, wherein, S4 specifically includes: S41, extract all relevant five-tuple knowledge from the knowledge graph, including five-tuples of manufacturer dataset table, platform dataset table, manufacturer dataset field, platform dataset field, manufacturer value domain, and platform value domain; S42, map the manufacturer dataset table and the platform dataset table, specifically including: Calculate the similarity between the manufacturer dataset table and the platform dataset table, the similarity calculation is based on table name, table annotation, and table structure; Select the best matching table mapping relationship according to similarity and confidence; S43, map the manufacturer dataset field and the platform dataset field, specifically including: Calculate the similarity between the manufacturer dataset field and the platform dataset field, the similarity calculation is based on field name, field annotation, field type, and field value domain; Select the best matching field mapping relationship according to similarity and confidence; S44, map the manufacturer value domain and the platform value domain, specifically including: Calculate the similarity between the manufacturer value domain and the platform value domain, the similarity calculation is based on the value range and enumeration value of the value domain; Select the best matching value domain mapping relationship according to similarity and confidence; S45, for the case where there are multiple candidate mappings, perform the following steps: Compare the confidence of the candidate mappings and select the mapping relationship with the highest confidence; If the confidence is the same, compare the reliability of the knowledge sources and prefer the mapping relationship with more reliable knowledge sources; If the knowledge sources are also the same, determine the final mapping relationship through manual review.

8. A cloud-computing-based medical information processing apparatus characterized by comprising: Include: The knowledge five-tuple construction module is deployed on the cloud computing platform and is used to construct a medical information processing knowledge five-tuple, the format of the five-tuple being: entity A-relation-entity B-confidence-knowledge source, wherein the entity A and the entity B are independently selected from the following entity types: standard data set entity, data table entity, table field entity, value domain entity, and quality control rule entity; the relation includes a mapping relation, an inheritance relation, an association relation, and a binding relation; The confidence is used to identify the accuracy weight of automatic matching or manual confirmation; The knowledge source is used to trace the knowledge origin; The knowledge extraction module is deployed on the cloud computing platform and is used to parse a medical institution database table structure and field annotation through a structured knowledge ETL engine to extract the five-tuple knowledge; and extract the five-tuple knowledge from medical industry standard documents, medical institution internal specifications, and other PDF files through an unstructured knowledge parsing engine; The knowledge storage module is deployed on the cloud computing platform and is used to store the extracted five-tuple according to the entity type to a graph database to construct a medical information processing knowledge graph; The mapping execution module is deployed on the cloud computing platform and is used to automatically execute the following mappings based on the five-tuple knowledge in the knowledge graph: mapping of a manufacturer data set table and a platform data set table; mapping of a manufacturer data set field and a platform data set field; mapping of a manufacturer value domain and a platform value domain; when there are multiple candidate mappings, selecting the mapping relation with the highest confidence; The feedback processing module is deployed on the cloud computing platform and is used to collect feedback data in the mapping execution process and adjust the confidence of the related five-tuple according to the feedback data.

9. An electronic device, comprising: The electronic device includes at least one processor and a memory connected in communication with the at least one processor; the memory stores computer program instructions executable by the at least one processor, and the computer program instructions are executed by the at least one processor to enable the at least one processor to execute the cloud computing-based medical information processing method according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, comprising: The non-transitory computer readable storage medium stores computer instructions, and the computer instructions enable the computer to execute the cloud computing-based medical information processing method according to any one of claims 1 to 7.

Images