Method and system for real-time update of medical knowledge base, and medium and device

Abstract

Provided in the present invention are a method and system for real-time update of a medical knowledge base, and a medium and a device. The method comprises: acquiring medical literature data in real time; analyzing the medical literature data, so as to obtain first structured data; performing correlation analysis on the first structured data and second structured data in a medical knowledge base, so as to obtain an analysis result; on the basis of the analysis result, determining whether to update the medical knowledge base; identifying the differences of the medical knowledge base before and after the update, so as to obtain a difference identification result; assigning a unique version number to each difference in the difference identification result, so as to obtain version update records; and on the basis of the version update records, using incremental hash table technology to perform data storage for each difference. The present application has a higher update efficiency without the risk of omission; and by means of difference analysis and by using incremental hash table technology to store difference data, the present application significantly reduces data storage space requirements, and also improves data access speed.

Description

Medical knowledge base real-time updating method, system, medium and device TECHNICAL FIELD

[0001] The present application relates to the technical field of database updating, in particular to a medical knowledge base real-time updating method, system, medium and device. BACKGROUND

[0002] In the current medical industry, the construction and application of knowledge base have become increasingly important, because the rapid development of medical knowledge and the growing amount of information require medical institutions to efficiently and accurately manage and use these knowledge. The core purpose of medical knowledge base is to centrally store and organize medical information, including disease diagnosis, treatment methods, drug information, and the latest medical research results, so that doctors and medical professionals can quickly obtain the required information and improve the efficiency and quality of diagnosis and treatment.

[0003] The prior art relies on manual retrieval and updating for medical knowledge base updating, which is inefficient and prone to missing important information.

[0004] SUMMARY

[0005] In view of the defects in the prior art, the purpose of the present application is to provide a medical knowledge base real-time updating method, system, medium and device.

[0006] According to the medical knowledge base real-time updating method provided by the present application, the medical literature data is obtained in real time, the medical literature data is analyzed to obtain first structured data, the first structured data and second structured data in the medical knowledge base are analyzed for correlation to obtain an analysis result, it is determined whether to update the medical knowledge base according to the analysis result, the differences before and after the update of the medical knowledge base are identified to obtain a difference identification result, a unique version number is assigned to each difference in the difference identification result to obtain a version update record, and the incremental hash table technology is used to store the data of each difference based on the version update record.

[0007] It is not difficult to understand that the present application is through structured analysis of medical literature data, and correlation calculation of structured data in the medical knowledge base, and updating of the medical knowledge base according to the correlation, compared with manual updating, the efficiency of the present application is higher, and there is no risk of omission. Moreover, the present application analyzes the differences before and after the update of the medical knowledge base, and assigns a version number to the updated medical knowledge base according to the difference analysis result, so as to reflect the update of the medical knowledge base, and facilitate understanding of the data differences between different medical knowledge base versions. Moreover, the present application uses incremental hash table technology to store the difference data, which significantly reduces the data storage space requirement, and improves the data access speed.

[0008] Further, the analyzing the medical literature data to obtain first structured data comprises: performing named entity recognition on the medical literature data to obtain named entities; filtering the named entities to obtain candidate entities; and grouping the candidate entities into different concept categories by using a clustering algorithm.

[0009] Further, the performing correlation analysis on the first structured data and second structured data in the medical knowledge base to obtain an analysis result comprises: extracting similarity feature dimensions from the first structured data and the second structured data; obtaining weight values corresponding to the similarity feature dimensions; and calculating a similarity degree according to the weight values and the sub-structured data.

[0010] Further, the determining whether to update the medical knowledge base according to the analysis result comprises: if the first structured data has a correlation with the second structured data in the medical knowledge base, updating the medical literature data to a corresponding position in the medical knowledge base; and if the first structured data does not have a correlation with the second structured data in the medical knowledge base, stopping updating the medical literature data to the medical knowledge base.

[0011] Further, the method for identifying differences between the medical knowledge base before and after the update comprises: generating Merkle trees for the medical knowledge base before and after the update, respectively; and identifying changed data items by comparing Merkle tree hash values of the medical knowledge base before and after the update, to generate a difference identification result.

[0012] It can be understood that, by constructing Merkle trees, the Merkle tree root hash values before and after the update of the medical knowledge base are compared, the changed data items are determined according to the hash difference, and the difference identification result is generated, thereby explicitly indicating the specific differences between the data versions, providing accurate difference data for version update and data storage, and effectively improving the accuracy and efficiency of data update.

[0013] Further, the method for assigning a unique version number to each difference in the difference identification result to obtain a version update record comprises: based on the difference identification result, creating a new version node for each difference, recording parent node information and change details of each version node, managing each version node in a linear update history chain manner, generating a version update record file, and recording update history and version change information of each data item.

[0014] It can be understood that, the application assigns a unique version number to each data change, ensures the consistency and traceability of version control, makes data rollback and version comparison operations feasible, and improves the flexibility and security of data management.

[0015] Further, the version-based update record is used to store each difference by using an incremental hash table technology, including: mapping the difference data of each version by using a key-value pair, the key being the version number of the data item, and the value being the difference content of the version relative to the previous version; adjusting the load factor and capacity parameters of the hash table to optimize the data retrieval and update efficiency, and generating the optimized storage difference data.

[0016] It can be understood that the application generates an optimized storage difference data set for different versions of the medical knowledge base, only containing the differences between the data versions, significantly reducing the data storage space requirement, and improving the data access speed. The optimized difference data set supports efficient data update and fast version switching, and provides a flexible and efficient data storage solution for the system.

[0017] According to the medical knowledge base real-time updating system provided by the application, the data processing and analysis module is used for analyzing the medical literature data to obtain first structured data, and the first structured data is analyzed in correlation with second structured data in the medical knowledge base to obtain an analysis result; the knowledge base integration module is used for determining whether to update the medical knowledge base according to the analysis result; the version control module is used for identifying the differences before and after the update of the medical knowledge base to obtain a difference identification result; a unique version number is assigned to each difference in the difference identification result to obtain a version update record; and the version update record is used to store each difference by using an incremental hash table technology.

[0018] According to the medical knowledge base real-time updating system provided by the application, the data processing and analysis module is used for analyzing the medical literature data to obtain first structured data, and the first structured data is analyzed in correlation with second structured data in the medical knowledge base to obtain an analysis result; the knowledge base integration module is used for determining whether to update the medical knowledge base according to the analysis result; the version control module is used for identifying the differences before and after the update of the medical knowledge base to obtain a difference identification result; a unique version number is assigned to each difference in the difference identification result to obtain a version update record; and the version update record is used to store each difference by using an incremental hash table technology.

[0019] According to the medical knowledge base real-time updating system provided by the application, the data processing and analysis module is used for analyzing the medical literature data to obtain first structured data, and the first structured data is analyzed in correlation with second structured data in the medical knowledge base to obtain an analysis result; the knowledge base integration module is used for determining whether to update the medical knowledge base according to the analysis result; the version control module is used for identifying the differences before and after the update of the medical knowledge base to obtain a difference identification result; a unique version number is assigned to each difference in the difference identification result to obtain a version update record; and the version update record is used to store each difference by using an incremental hash table technology.

[0020] Compared with the prior art, the application has the following beneficial effects:

[0021] Compared with manual updating, the efficiency of the application is higher, and the risk of omission is avoided. Moreover, the application performs difference analysis on the medical knowledge bases before and after updating, and assigns a version number to the updated medical knowledge base according to the difference analysis result, so that the updating of the medical knowledge base can be reflected, and the difference between different versions of the medical knowledge base can be understood. Moreover, the application uses incremental hash table technology to store the difference data, which significantly reduces the data storage space requirement and improves the data access speed. BRIEF DESCRIPTION OF DRAWINGS

[0022] Other features, objects and advantages of the application will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings:

[0023] Fig. 1 is a flow chart of a medical knowledge base real-time updating method according to the application;

[0024] Fig. 2 is a structural diagram of a Merkle tree according to the application;

[0025] Fig. 3 is a flow chart of a first structured data acquisition process according to the application;

[0026] Fig. 4 is a deep network model diagram according to the application;

[0027] Fig. 5 is a flow chart of a similarity calculation method according to the application;

[0028] Fig. 6 is a system architecture diagram of a medical knowledge base real-time updating system according to the application. DETAILED DESCRIPTION

[0029] The application will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the application, but do not limit the application in any form. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the application. These all belong to the protection scope of the application.

[0030] The application discloses a medical knowledge base real-time updating method, comprising: analyzing medical literature data to obtain first structured data, and determining the relevance of the first structured data and second structured data in a medical knowledge base; integrating the first structured data meeting a preset condition into the medical knowledge base to obtain an updated medical knowledge base; identifying the difference between versions before and after updating the medical knowledge base to obtain a difference identification result; assigning a unique version number to each difference in the difference identification result to obtain a version update record; and storing each difference based on the version update record using incremental hash table technology.

[0031] It is not difficult to understand that, compared with manual updating, the efficiency of the present application is higher and there is no risk of omission. Moreover, the present application performs difference analysis on the medical knowledge base before and after updating, and assigns a version number to the updated medical knowledge base according to the difference analysis result, so as to reflect the updating of the medical knowledge base and facilitate understanding of the difference between different versions of the medical knowledge base. Moreover, the present application uses incremental hash table technology to store the difference data, which significantly reduces the data storage space requirement and improves the data access speed.

[0032] Referring to FIG. 1, a medical knowledge base real-time updating method includes:

[0033] Step 1: Real-time acquisition of medical literature data.

[0034] The present application realizes real-time monitoring and acquisition of medical data on the Internet through automatic scripts, for example, real-time monitoring of various medical data sources through RSS subscription and API calling. The medical data sources include but are not limited to: authoritative medical literature including medical data, medical guidelines, papers, etc.; clinical databases including electronic medical records, images, tests, and a large number of data sets generated by professional clinical business systems; and various Internet information including medical information, question and answer data, etc. Of course, it can also be other data sources, which are not limited here.

[0035] Step 2: Analyzing the medical literature data to obtain first structured data.

[0036] In some embodiments, analyzing the medical literature data to obtain the first structured data includes: performing named entity recognition on the medical literature data based on a deep learning model to obtain named entities, grouping the named entities into different concepts through a clustering algorithm according to the obtained named entities, performing relationship extraction on different concepts to obtain semantic relationships between the concepts, and obtaining the first structured data according to the concepts and the semantic relationships between the concepts.

[0037] It is not difficult to understand that, by performing named entity recognition on the medical literature data and classifying the obtained named entities through a clustering algorithm, the present application can summarize and extract key information from the massive content of the medical literature data, thereby facilitating subsequent comparison of the medical literature data with the knowledge base. By performing relationship extraction on the concepts and further strengthening the structure division of the medical literature data according to the semantic relationships between the concepts, the present application can more clearly understand the logical structure of the medical literature data, thereby facilitating the final obtaining of the first structured data.

[0038] Step 3: Correlation analysis is performed on the first structured data and the second structured data in the medical knowledge base, and an analysis result is obtained.

[0039] In some embodiments, the correlation analysis on the first structured data and the second structured data in the medical knowledge base can be performed according to the similarity between the first structured data and the second structured data in the medical knowledge base, and an analysis result is obtained. The analysis result includes: the first structured data and the second structured data in the medical knowledge base have correlation, the first structured data and the second structured data in the medical knowledge base do not have correlation.

[0040] For example, similarity calculation is performed on the first structured data and the second structured data in the medical knowledge base, and a value of the similarity is obtained. If the value of the similarity is greater than a first preset value, it is determined that the first structured data and the second structured data in the medical knowledge base have correlation. If the value of the similarity is not greater than the first preset value, it is determined that the first structured data and the second structured data in the medical knowledge base do not have correlation. The first preset value can be adjusted according to actual conditions, for example, it can be 70%.

[0041] Step 4: According to the analysis result, it is determined whether to update the medical knowledge base.

[0042] As can be seen from step 3, according to the analysis result, if the first structured data and the second structured data in the medical knowledge base have correlation, it is considered that the medical literature data and the content of the medical knowledge base have relevance, and the medical literature data needs to be updated to the corresponding position in the medical knowledge base. If the first structured data and the second structured data in the medical knowledge base do not have correlation, it is considered that the medical literature data and the content of the medical knowledge base do not have relevance, and the medical literature data does not need to be updated to the medical knowledge base.

[0043] Step 5: Differences before and after updating the medical knowledge base are identified, and a difference identification result is obtained.

[0044] In a specific embodiment, identifying the differences before and after updating the medical knowledge base to obtain a difference identification result includes: generating a Merkle tree for the medical knowledge base before and after updating respectively, and identifying the differences before and after updating the medical knowledge base by executing a Merkle tree algorithm, including calculating a hash value for each data item, using a SHA-256 algorithm to hash the data block, identifying the changed data item by comparing the hash value of the current knowledge base data with the hash value of the knowledge base data before updating, and generating a difference identification result.

[0045] In the embodiments of the present application, how to perform difference identification is further illustrated in combination with the structure of the Merkle tree. Referring to FIG. 2, the Merkle tree is a hierarchical data structure that represents a group of data as a cryptographic hash tree. The construction of the Merkle tree is a recursive process that starts from the bottom data block and continuously calculates the hash value of the parent node upwards until the root node is reached. Starting from the bottom, D0, D1, D2, and D3 are the data contained in the leaf nodes, that is, the value of the leaf node. Continuing to look upwards, N0, N1, N2, and N3 are leaf nodes, which are the hash values obtained by performing hash operation on the data, that is, D0, D1, D2, and D3. Continuing to look upwards, N4 and N5 are non-leaf nodes, which are the hash values obtained by performing hash operation on N0 and N1 and N2 and N3, respectively. The root node is the hash value obtained by performing hash operation on N4 and N5, which is the root hash of the Merkle tree. When identifying the difference between the medical knowledge base before and after the update, the root hash values of the Merkle trees before and after the update of the medical knowledge base, the hash values corresponding to the non-leaf nodes, and the hash values corresponding to the leaf nodes are compared in sequence, so as to quickly identify the nodes with different hash values, determine the changed data items, generate the difference identification result, and clearly indicate the specific difference between the medical knowledge base before and after the update. The accurate difference data is provided for subsequent version update and data storage, and the accuracy and efficiency of data update are effectively improved.

[0046] Step 6: Assign a unique version number to each difference in the difference identification result to obtain a version update record.

[0047] Specifically, based on the difference identification result, a new version node is created for each difference, the parent node information and change details of each node are recorded, the linear update history chain is used to manage each version node, the version update record file is generated, and the update history and version change information of each data item are recorded. For example, when medical literature data is updated into the medical knowledge base, multiple data items may have been changed, and a unique version number needs to be assigned to each data item.

[0048] Creating a new version node for each difference includes: based on the difference identification result, performing a version control process to create a new version node for each identified difference, recording the change details of the data item and the parent node information, and ensuring the completeness and traceability of the version history. The linear update history chain is used to manage each version node, a unique version number is assigned to each data change, the consistency and traceability of version control are ensured, the version update record file is generated, and the update history and version change information of each data item are recorded, so that data rollback and version comparison operations are feasible, and the flexibility and security of data management are improved.

[0049] Step 7: Based on the version update record, incremental hash table technology is used to store the data of each difference.

[0050] Specifically, the difference data of each version is mapped to a key-value pair, with the key being the version number of the data item and the value being the difference content of the version relative to the previous version. The load factor and capacity parameters of the hash table are adjusted to optimize data retrieval and update efficiency, and the optimized stored differential data is generated. Based on the version update record, incremental hash table technology is used for data storage. The difference data is mapped to a key-value pair, with the key being the version number and the value being the difference content. By adjusting the load factor and capacity parameters of the hash table, the data retrieval and update efficiency is optimized, and the optimized stored differential data is generated.

[0051] Based on the version update record, incremental hash table technology is used for data storage. The differential data is stored in a key-value pair format, with the key being the version number of the data item and the value being the difference content of the version relative to the previous version. By adjusting the load factor and capacity parameters of the hash table, the data storage structure and retrieval efficiency are optimized. The optimized stored differential data set only contains the differences between data versions, significantly reducing the data storage space requirement while improving the data access speed. The optimized differential data set supports efficient data update and fast version switching, providing a flexible and efficient data storage solution for the system.

[0052] Specifically, referring to FIG. 3, the process of obtaining the first structured data by analyzing the medical literature data in step 2 includes:

[0053] Step 201: Based on a deep learning model, named entity recognition is performed on the medical literature data to obtain named entities;

[0054] Step 202: According to the obtained named entities, a clustering algorithm is used to classify the named entities into different concepts;

[0055] In one specific embodiment, the method of classifying named entities into different concepts in step 202 is as follows:

[0056] The named entities are classified into corresponding concept categories through a clustering algorithm such as K-means. Specifically, the semantic representation features and co-occurrence relationship features of the named entities are comprehensively utilized during clustering, and the two features are weighted and fused. The semantic representation features are obtained by extracting the embedding vectors of the named entities through BERT, reflecting the context semantic information of the named entities. The co-occurrence relationship features describe the common occurrence patterns of the named entities in the text, such as named entities appearing in the same sentence or paragraph may belong to similar concepts. For example, the named entities "gastric cancer", "liver cancer", and "lung cancer" are classified into the "cancer" concept category.

[0057] Specifically, the semantic representation features and the contribution relationship features are further described in combination with specific models. The semantic representation features are vectors that map words to a high-dimensional space using a pre-trained BERT, which can capture the semantic similarity and relationship between words. The co-occurrence relationship features are obtained using Term Frequency-Inverse Document Frequency (TF-IDF), which measures the importance of a word in a document, i.e., the product of the frequency of the word in the document and the general importance of the word in the document set. The semantic representation features and the co-occurrence relationship features are fused to form comprehensive feature representation. This can be achieved by simple feature concatenation, weighted summation or more complex feature fusion methods. Here, no limitation is made.

[0058] Step 203: Relationship extraction is performed on different concepts to obtain semantic relationships between the concepts, and first structured data is obtained based on the concepts and the semantic relationships between the concepts.

[0059] In one specific embodiment, the relationship extraction method for concepts in step 203 includes:

[0060] Dependence analysis based on transition is performed on the concepts to analyze the dependence relationships between the concepts and find core concepts; recursive search is performed on all concepts parallel to the core concepts to form a predicate list; for each predicate, its subject and object are found based on the dependence relationship to form an information tuple; for an information tuple without a subject, a parent tuple parallel to its predicate and having a subject is found, and if the parent tuple exists, the subject of the information tuple is set as the subject of the parent tuple, and if no parent tuple is found, the information tuple is not processed; and the processing result is returned to obtain the syntactic relationships between the concepts.

[0061] In the concept relationship extraction process, for isolated concepts that cannot be extracted through the relationships between concept words, semantic relationships are established between the isolated concepts and other concepts based on their semantic similarity. Finally, first structured data is obtained based on the obtained concepts and the semantic relationships between the concepts.

[0062] In one specific embodiment, the method for performing named entity recognition on medical literature data based on a deep learning model in step 201 includes: determining the probability of different labels corresponding to each character of the medical literature data using a deep learning model, determining the label of each character of the medical literature data based on the probability of different labels corresponding to each character of the medical literature data, determining the label sequence of the medical literature data, and determining the named entity of the medical literature data based on the determined label sequence of the medical literature data.

[0063] Taking BIO labeling as an example, the different labels corresponding to each character include B, I and O, wherein B and I collectively represent an entity, B represents the start position of a named entity, I represents a position within a named entity, and O represents a non-entity or a non-target category entity. For example, "viruses can cause colds" is labeled as:

[0064] Table 1 BIO annotation example

[0065] As shown in FIG. 4, a specific deep learning model structure is disclosed for named entity recognition of medical literature data. Specifically, the deep learning model can include an embedding layer 101 and a neural network layer 102, wherein the neural network layer 102 includes a feature extraction layer 1021 and a fully connected layer 1022.

[0066] The embedding layer 101 is configured to convert each character of the medical literature data into a word vector and a word vector to which each character belongs, and concatenate the word vector and the word vector to obtain a concatenated vector.

[0067] Specifically, the embedding layer 101 is configured to obtain input medical literature data, convert each character of the medical literature data into a word vector and a word vector to which each character belongs. The adopted manner can be word embedding such as word2vec, glovec, fasttext, etc., which maps each word to a real number vector, and these vectors are semantically related.

[0068] The feature extraction layer 1021 is configured to determine a concatenated feature vector by performing feature extraction on the concatenated vector.

[0069] In some embodiments, since medical text contains a large number of professional terms and proper nouns and other special entities, context information needs to be relied on to accurately extract these entities. Therefore, the neural network layer 102 of the present application can adopt a Bi-directional Long Short-Term Memory (BiLSTM) network. The BiLSTM network can fully consider the context semantic information of the electronic health record data, avoid the long distance dependency problem of the medical text, so as to obtain more information of the concatenated feature vector, thereby improving the accuracy of the fully connected layer 1022 in determining the probability of each character of the electronic health record data corresponding to different labels.

[0070] The fully connected layer 1022 is configured to obtain the probability of each character of the electronic health record data corresponding to different labels according to the concatenated feature vector.

[0071] Specifically, the full connection layer 1022 maps the spliced feature vector to a probability distribution corresponding to each label through an activation function, and then normalizes the probability distribution of different labels corresponding to each character to generate the probability of different labels corresponding to each character. According to the combination of the label with the maximum probability of each character, the label sequence corresponding to the electronic health record data can be obtained, and the named entity recognized from the electronic health record data can be obtained according to the label sequence.

[0072] In a specific embodiment, referring to FIG. 5, the first structured data and the second structured data similarity calculation method includes:

[0073] Step S301, extracting similarity feature dimensions from the first structured data and the second structured data. For example, for the literature related to chronic obstructive pulmonary disease, the dimensions of “age, gender, BMI, number of acute exacerbations in one year, smoking history, smoking pack years, recent cough, recent sputum, recent chest tightness, recent wheezing, recent shortness of breath, past shortness of breath, nocturnal paroxysmal dyspnea, bilateral lower extremity edema, palpitations, blood eosinophil count, sputum eosinophil count, and comorbidities (multiple)” are selected as similarity feature dimensions, and the structured data under these dimensions in the first structured data and the second structured data can be extracted as sub-structured data.

[0074] Step S302, obtaining the weight value corresponding to the similarity feature dimension.

[0075] The determined similarity feature dimensions are used as similarity comparison nodes, and the number of similarity feature dimensions is generally multiple. In this step, each node can be assigned a corresponding weight value according to the size of the influence factor of each similarity feature dimension on the similarity.

[0076] Step S303, calculating the similarity according to the weight value and the sub-structured data. In this step, the similarity can be calculated according to the weight value assigned to each similarity feature dimension and the sub-structured data under each similarity feature dimension, for example, the similarity can be calculated by weighted summation of each sub-structured data. It can also be: obtaining the intersection value and the union value of the sub-structured data of the first structured data and the second structured data; then, after obtaining the weight value assigned to each similarity feature dimension, the weighted intersection value is obtained by weighting calculation using the intersection value of each sub-structured data and the corresponding weight value, the weighted union value is obtained by weighting calculation using the union value of each sub-structured data and the weight value, and finally the ratio of the weighted intersection value to the weighted union value is taken as the similarity.

[0077] The following further illustrates how to use the ratio of the weighted intersection value and the weighted union value as the similarity. Assuming that each sub-structured data of the first structured data is represented as set A, and each sub-structured data of the second structured data is represented as set B, the intersection of set A and set B can be calculated first, and the weighted sum of each item in the intersection is obtained to obtain the weighted intersection value. Then, the union of set A and set B is calculated, and the weighted sum of each item in the union is obtained to obtain the weighted union value. Finally, the similarity of the first structured data and the second structured data can be obtained by dividing the weighted intersection value by the weighted union value, which can be represented by the following formula:

[0078] wherein S represents the similarity, wi represents the product of the i-th data in the intersection of set A and set B and the corresponding weight value, and wj represents the product of the j-th data in the union of set A and set B and the corresponding weight value. The similarity of the structured data of each electronic medical record can be accurately calculated by the weighted intersection value and the weighted union value of each sub-structured data of each electronic medical record in the embodiment, which provides a data basis for accurately calculating the similarity between electronic medical records.

[0079] Referring to FIG. 6, a real-time updating system of a medical knowledge base is used to execute the above-mentioned structured processing method of electronic health records. The system comprises:

[0080] a data processing and analysis module for analyzing medical literature data to obtain first structured data and the relevance of the first structured data and second structured data in the medical knowledge base;

[0081] a knowledge base integration module for integrating the first structured data whose relevance meets a preset condition into the medical knowledge base to obtain a new version of the medical knowledge base;

[0082] a version control module for identifying the difference between the versions of the medical knowledge base before and after updating to obtain a difference identification result, assigning a unique version number to each difference in the difference identification result to obtain a version update record, and storing each difference based on the version update record using an incremental hash table technology.

[0083] In other embodiments, the real-time updating system of the medical knowledge base further comprises a user interface module, a version control system, a user feedback and quality control module, a system integration and coordination module, and a deployment and maintenance module.

[0084] The user interface module is used to provide a user interface for users to view the update history, differences, and knowledge base status. The version control system is used to track the change history of the knowledge base, recording the detailed differences of each update. The differential storage and version comparison algorithm is implemented to provide the historical version recovery function. The version information and change records are transmitted to the user interface module for user query. The user feedback and learning module is used to collect user feedback, conduct quality control, ensure the accuracy and reliability of the knowledge base information, and implement the user feedback mechanism to verify and correct the errors or omissions reported by users. The system integration and testing module is used to ensure the coordinated work of the system modules, conduct system testing and performance evaluation. Through the standardization of module interfaces and data synchronization mechanism, integration testing and stress testing are conducted, and the system integration and testing results are fed back to each module to guide system optimization and upgrading. The deployment and maintenance module is responsible for the deployment, monitoring and maintenance of the system to ensure stable operation of the system. Through the integrated monitoring system, the system performance indicators are tracked in real time, and the automatic deployment and rollback mechanism is realized. At the same time, system operation data and maintenance logs are collected for continuous system improvement.

[0085] It is to be understood that the embodiments of the present application can be implemented in hardware, software, or a combination thereof. The software can be stored in a volatile or non-volatile storage device like a ROM, whether or not the data is deletable or rewritable, in a memory like a RAM, a memory chip, a device, or an integrated circuit, or in a storage medium on which data can be optically or magnetically recorded and from which data can be read by a machine (e.g., a computer), such as a compact disc (CD), a digital versatile disc (DVD), a magnetic disk, or a magnetic tape. Also, the embodiments of the present application can be implemented in a computer or a portable terminal having a controller and a memory, and the memory is an example of a non-transitory computer-readable storage medium suitable for storing one or more programs including commands for implementing the embodiments of the present application. Therefore, the present application includes a program having a code for implementing an apparatus or a method defined by the claims, and a machine-readable non-transitory storage medium storing the program. The program can be transferred electrically through a medium such as a communication signal transmitted via a wired or wireless connection, and the program and its equivalents are included in the present application.

[0086] The device can receive a program and store the program from a program providing device connected through a cable or wirelessly. The program providing device can include a program including commands for implementing the embodiments of the present application, a memory for storing information required for the embodiments of the present application, a communication module for communicating with the mobile device through a wired or wireless connection, and a controller for automatically or according to a request of the device, transferring the program to the device.

[0087] Those skilled in the art know that, in addition to implementing the system provided by the present application and each device, module and unit thereof in the form of pure computer readable program code, the system provided by the present application and each device, module and unit thereof can also be implemented in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers and embedded microcontrollers, etc. by logically programming the method steps to achieve the same functions. Therefore, the system provided by the present application and each device, module and unit thereof can be considered as a hardware component, and the devices, modules and units included therein for achieving various functions can also be considered as structures within the hardware component; the devices, modules and units for achieving various functions can also be considered as both software modules for achieving methods and structures within hardware components.

[0088] The specific embodiments of the present application are described above. It needs to be understood that the present application is not limited to the specific embodiments described above, and various changes or modifications can be made by those skilled in the art within the scope of the claims, which does not affect the essential content of the present application. The embodiments of the present application and the features in the embodiments can be combined with each other in any manner without conflict.

Claims

1. A medical knowledge base real-time updating method, characterized by, The method comprises the following steps: real-time acquisition of medical literature data; analysis of the medical literature data to obtain first structured data; correlation analysis of the first structured data and second structured data in a medical knowledge base to obtain an analysis result; determination of whether to update the medical knowledge base according to the analysis result; identification of differences before and after the update of the medical knowledge base to obtain a difference identification result; allocation of a unique version number to each difference in the difference identification result to obtain a version update record; and data storage of each difference based on the version update record by using an incremental hash table technology.

2. The medical knowledge base real-time updating method according to claim 1, characterized in that, The analysis of the medical literature data to obtain the first structured data comprises the following steps: named entity recognition of the medical literature data to obtain named entities; screening of candidate entities from the named entities, and clustering of the candidate entities into different concept categories by using a clustering algorithm; acquisition of semantic relationships between the candidate entities based on the concept categories to obtain the first structured data.

3. The medical knowledge base real-time updating method according to claim 1, wherein, The correlation analysis of the first structured data and the second structured data in the medical knowledge base to obtain the analysis result comprises the following steps: extraction of similarity feature dimensions from the first structured data and the second structured data; acquisition of weight values corresponding to the similarity feature dimensions; calculation of a similarity degree according to the weight values and the sub-structured data.

4. The medical knowledge base real-time updating method according to claim 3, wherein, The determination of whether to update the medical knowledge base according to the analysis result comprises the following steps: if the first structured data has a correlation with the second structured data in the medical knowledge base, the medical literature data is updated to a corresponding position in the medical knowledge base; and if the first structured data does not have a correlation with the second structured data in the medical knowledge base, the updating of the medical literature data to the medical knowledge base is stopped.

5. The medical knowledge base real-time updating method of claim 1, wherein, The method for identifying the differences before and after the update of the medical knowledge base comprises the following steps: generation of Merkle trees for the medical knowledge base before and after the update, respectively; identification of changed data items by comparing the Merkle tree hash values of the medical knowledge base before and after the update to generate a difference identification result. The allocation of a unique version number to each difference in the difference identification result to obtain the version update record comprises the following steps:

6. The medical knowledge base real-time updating method of claim 1, wherein, creation of a new version node for each difference based on the difference identification result, recording of parent node information and change details of each version node, management of each version node in a linear update history chain, generation of a version update record file, and recording of update history and version change information of each data item. The data storage of each difference based on the version update record by using the incremental hash table technology comprises the following steps:

7. The medical knowledge base real-time updating method of claim 1, wherein, key-value pair mapping of difference data of each version, with the key being a version number of the data item and the value being difference content of the version relative to a previous version; adjustment of a load factor and a capacity parameter of the hash table to optimize data retrieval and update efficiency, and generation of optimized storage difference data. The method comprises the following steps:

8. A medical knowledge base real-time updating system, characterized by comprising: analysis of medical literature data by a data processing and analysis module to obtain first structured data; correlation analysis of the first structured data and second structured data in a medical knowledge base to obtain an analysis result; determination of whether to update the medical knowledge base according to the analysis result by a knowledge base integration module; identification of differences before and after the update of the medical knowledge base to obtain a difference identification result by a version control module. ​ A unique version number is assigned to each difference in the difference identification result to obtain a version update record; and based on the version update record, an incremental hash table technique is used to store data of each difference.

9. A computer readable storage medium storing a computer program, characterized in that, The computer program, when executed by a processor, implements the steps of the medical knowledge base real-time updating method of any one of claims 1 to 7.

10. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The computer program, when executed by a processor, implements the steps of the medical knowledge base real-time updating method of any one of claims 1 to 7.

Images