Knowledge Graph-based Sleep Disorder Attribution Analysis Method, Device and System

Abstract

The present invention discloses a method, device and system for sleep disorder attribution analysis based on a knowledge graph. By constructing a domain knowledge graph, extracting subgraphs from the graph, constructing structural prior knowledge, and finally performing attribution analysis using a trained Bayesian network, it can effectively utilize expert experience, reduce the requirement of the model for the sample size, improve the performance of the model, accelerate the model training speed, and obtain effective analysis results by combining expert experience with model training in the case of lack of big data on mental diseases or sleep disorders.

Description

[0001] This application is a divisional application of a Chinese invention patent application with an application date of August 12, 2021, an application number of CN202110921652.9, and an invention title of "Method, Device and System for Attributing Analysis of Sleep Disorders Based on Knowledge Graph". Technical Field

[0002] The present invention relates to the field of artificial intelligence, and in particular, to a method, device and system for attributing analysis of sleep disorders based on a knowledge graph. Background Art

[0003] With the accelerating pace of modern society and increasing life pressure, mental health has increasingly become the focus of attention of working people. Diseases caused by mental or psychological factors such as depression and neurasthenia are increasingly troubling many people. Establishing a model through artificial intelligence algorithms for the examination results of patients by means of big data can assist doctors or medical workers to provide more scientific judgments. How to effectively utilize this data and provide better decisions is a scientific problem that needs to be solved urgently. As the factors affecting disease judgment continue to increase and the changes in indicators become normal, how to discover the potential factors promoting the growth of indicators is becoming a difficult problem. Summary of the Invention

[0004] Aiming at the above defects, the technical problem to be solved by the present invention is how to make better use of the big data of various factors affecting mental or sleep disorders and the increasing dimensions for more scientific judgments.

[0005] Aiming at the above defects, the purpose of the present invention is to provide a method, system, electronic device, computer storage medium and program product for attributing analysis of sleep disorders based on a knowledge graph.

[0006] According to one aspect of the embodiments of the present specification, a method for attributing analysis of sleep disorders based on a knowledge graph is provided for the server side. Through an algorithm model, entity extraction is performed on the collected data, the knowledge is stored using a graph database, and a knowledge graph is constructed in combination with the experience of business experts.

[0007] According to the knowledge graph and the input information, relevant subgraphs are extracted, a structural prior is constructed, and in combination with training samples, a Bayesian network model is learned and established.

[0008] The learned and established Bayesian network model is used for attribution analysis.

[0009] Preferably, the algorithm model includes natural language processing, deep learning and knowledge graph technology.

[0010] Preferably, entity extraction includes relation extraction, event extraction, entity disambiguation, knowledge fusion and knowledge processing.

[0011] Preferably, the structural prior is constructed by extracting the subgraph structure, directly constructing the Bayesian network structure parameter distribution, and then combining the samples to jointly learn the Bayesian network structure.

[0012] Preferably, a structural prior is constructed to count the frequency of variables in the sample and the frequency between variables, and the average frequency of the variables and the average frequency between variables are calculated. According to the subgraph structure, the parent node is used as the tail node and the child node is used as the head node. The frequency and the average frequency are used to obtain the probability distribution between nodes. According to the subgraph structure, the probability distribution between nodes is repeatedly constructed, and the obtained subgraph probability distribution is used as a structural prior parameter. The Bayesian network structure is obtained in combination with sample learning.

[0013] Preferably, the structure prior is constructed by adding a penalty factor to the scoring function so that the prior structure is integrated into the posterior structure.

[0014] The present invention provides a sleep disorder attribution analysis method based on knowledge graph, which is applied to the Internet medical platform, collects medical diagnosis and examination data input by users, extracts entities from the data through an algorithm model, uses a graph database to store knowledge, and constructs a knowledge graph in combination with the experience of doctors and experts;

[0015] According to the knowledge graph and input information, relevant subgraphs are extracted, structural priors are constructed, and combined with training samples, a Bayesian network model is learned and established;

[0016] The Bayesian network model established through learning is used for attribution analysis to form analysis results.

[0017] Preferably, the medical diagnosis and examination data includes text data and image data.

[0018] Preferably, entity extraction includes extraction of relationships between sleep disorders and medical diagnosis and examination data, extraction of user behavior events, expert experience entity disambiguation, disease and symptom knowledge graph knowledge fusion and knowledge processing.

[0019] Preferably, the extracted subgraph uses a graph neural network model to predict the relationship between nodes in the graph and mine more causal relationships.

[0020] Preferably, the method combines expert experience to construct a knowledge graph of related business fields, extracts relevant subgraphs based on images, uses a graph neural network model on the subgraphs to predict the relationship between nodes, mines more causal relationships, constructs a priori distribution of Bayesian network structures, and combines samples to learn the Bayesian network.

[0021] The present invention provides a sleep disorder attribution analysis system based on knowledge graph, including a server, a client and an Internet medical platform.

[0022] The user submits medical diagnosis and examination data through the client,

[0023] The Internet medical platform collects medical diagnosis and examination data input by users. The server side extracts entities from the data through an algorithm model, stores the knowledge using a graph database, and constructs a knowledge graph by combining the experience of doctors and experts.

[0024] According to the knowledge graph and the input information, relevant subgraphs are extracted, a structural prior is constructed, and combined with training samples to learn and establish a Bayesian network model.

[0025] Use the learned Bayesian network model for attribution analysis to form an analysis result.

[0026] Preferably, entity extraction includes the extraction of the relationship between sleep disorders and medical diagnosis and examination data, the extraction of user behavior events, the disambiguation of expert experience entities, the knowledge fusion and knowledge processing of disease and symptom knowledge graphs.

[0027] Preferably, to construct the structural prior, the frequency of variables and the frequency between variables in the sample are counted, the average frequency of variables and the average frequency between variables are calculated. According to the subgraph structure, the parent node is used as the tail node and the child node is used as the head node, and the probability distribution between nodes is obtained using the frequency and average frequency. According to the subgraph structure, the probability distribution between nodes is repeatedly constructed, and the obtained subgraph probability distribution is used as the structural prior parameter, and combined with the sample to learn the Bayesian network structure.

[0028] The present invention provides a computer-readable storage medium, on which a computer program / instructions are stored, characterized in that when the computer program / instructions are executed by a processor, the steps of the above method are implemented.

[0029] The present invention provides a computer program product, including a computer program / instructions, characterized in that when the computer program / instructions are executed by a processor, the steps of the above method are implemented.

[0030] The present invention provides an electronic device, including:

[0031] A processor; and

[0032] A memory configured to store computer-executable instructions, and when the executable instructions are executed, the processor is caused to perform the following operations:

[0033] Through an algorithm model, extract entities from the collected data, store the knowledge using a graph database, and construct a knowledge graph by combining the experience of business experts;

[0034] According to the knowledge graph and the input information, extract relevant subgraphs, construct a structural prior, and combine with training samples to learn and establish a Bayesian network model;

[0035] Perform attribution analysis using the Bayesian network model established through learning.

[0036] The present invention provides an electronic device, including:

[0037] A processor; and

[0038] A memory configured to store computer-executable instructions, and the executable instructions, when executed, cause the processor to perform the following operations:

[0039] Collect medical diagnostic examination data input by a user, perform entity extraction on the data through an algorithm model, store knowledge using a graph database, and construct a knowledge graph in combination with the experience of doctor experts;

[0040] According to the knowledge graph and the input information, extract relevant subgraphs, construct a structural prior, and in combination with training samples, learn to establish a Bayesian network model;

[0041] Perform attribution analysis using the established Bayesian network model through learning to form an analysis result.

[0042] The present invention can effectively utilize expert experience, reduce the requirements of the model for the sample size, improve the performance of the model, accelerate the model training speed, and obtain an effective analysis result by combining expert experience with model training in the case of a lack of big data on mental diseases or sleep disorders. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required to be used in the embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0044] Figure 1 Shows a schematic framework diagram of an embodiment of the sleep disorder attribution analysis method based on a knowledge graph according to the present invention;

[0045] Figure 2 Shows a schematic framework diagram of another embodiment of the sleep disorder attribution analysis method based on a knowledge graph according to the present invention;

[0046] Figure 3 Shows a schematic flowchart of an embodiment of the sleep disorder attribution analysis method based on a knowledge graph according to the present invention;

[0047] Figure 4 Shows a schematic flowchart of another embodiment of the sleep disorder attribution analysis method based on a knowledge graph according to the present invention;

[0048] Figure 5 Shows a schematic flowchart of another embodiment of the sleep disorder attribution analysis method based on a knowledge graph according to the present invention. Detailed Implementation Modes

[0049] The features and exemplary embodiments of various aspects of the present invention will be described in detail below. To make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only configured to explain the present invention and are not configured to limit the present invention. For those skilled in the art, the present invention can be implemented without some of these specific details. The following description of the embodiments is only intended to provide a better understanding of the present invention by showing examples of the present invention.

[0050] It should be noted that, in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise", or any other variant thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or device including a series of elements not only includes those elements but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article, or device. Without further limitation, elements defined by the statement "including..." do not exclude the presence of additional identical elements in the process, method, article, or device including the said elements.

[0051] As Figure 1 shown, a method for sleep disorder attribution analysis based on a knowledge graph provided by an embodiment of this specification is used for the server side. Through an algorithm model, entity extraction is performed on the collected data, the knowledge is stored using a graph database, and a knowledge graph is constructed in combination with the experience of business experts;

[0052] According to the knowledge graph and input information, relevant subgraphs are extracted, a structural prior is constructed, and in combination with training samples, a Bayesian network model is learned and established;

[0053] The learned and established Bayesian network model is used for attribution analysis.

[0054] Causal judgment is a process of drawing conclusions about causal relationships based on the conditions under which an effect occurs.

[0055] A knowledge graph describes concepts, entities, and the relationships between them in the objective world in a structured form, expressing the information of Internet medicine in a form closer to the human cognitive world, and having the ability to better organize, manage, and understand the massive information in Internet medical big data.

[0056] Graph neural networks are typified by the GNN network structure model. In probability and information theory, mutual information (MI) measures the degree of mutual dependence between two random variables.

[0057] Conditional mutual information (CMI) measures the degree of direct or indirect non - linear dependence between variables X and U given variable Z among three variables X, Y, and Z.

[0058] Partial mutual information (PMI) has basically the same principle as CMI. The difference is that X and Y have partial independence rather than complete conditional independence given Z.

[0059] In some embodiments, the algorithm model includes natural language processing, deep learning, and knowledge graph technology.

[0060] In some embodiments, entity extraction includes relation extraction, event extraction, entity disambiguation, knowledge fusion, and knowledge processing.

[0061] In some embodiments, the structural prior is constructed by extracting sub - graph structures, directly constructing the Bayesian network structure parameter distribution, and then jointly learning the Bayesian network structure in combination with samples.

[0062] As Figure 2 shown, an attribution analysis method for sleep disorders based on a knowledge graph provided by an embodiment of this specification uses an algorithm model to perform entity extraction on the collected data, stores the knowledge using a graph database, and constructs a knowledge graph in combination with the experience of business experts;

[0063] According to the knowledge graph and input information, relevant sub - graphs are extracted, a structural prior is constructed, and combined with training samples, a Bayesian network model is learned and established;

[0064] The learned Bayesian network model is used for attribution analysis;

[0065] In the process of learning the Bayesian network structure, according to the relevant theories of information theory, the linear and non - linear causal relationships between variables are better measured. In this embodiment, partial mutual information (PMI) is added to accurately measure the linear and non - linear causal relationships between variables.

[0066] Let X, Y, and Z be three random variables. According to the relevant knowledge of information theory, the definitions of mutual information and conditional mutual information are as follows:

[0067]

[0068]

[0069] In the formula, the definition of partial mutual information is as follows:

[0070]

[0071] Among them,

[0072]

[0073]

[0074] By adding the above-mentioned partial mutual information (PMI), the serious overestimation and underestimation problems of MI and CMI when measuring linear and non-linear causal relationships between variables can be solved.

[0075] Such as Figure 3 As shown, an embodiment of the present specification provides a method for sleep disorder attribution analysis based on a knowledge graph, including:

[0076] S101. Extract entities from the collected data through an algorithm model;

[0077] S102. Store knowledge using a graph database;

[0078] S103. Construct a knowledge graph by combining the experience of business experts;

[0079] S104. Extract relevant subgraphs according to the knowledge graph and input information, and construct a structural prior;

[0080] S105. Combine training samples to learn and establish a Bayesian network model;

[0081] S106. Perform attribution analysis using the learned Bayesian network model.

[0082] In a specific example, the structural prior is constructed to count the frequencies of variables and the frequencies between variables in the sample, calculate the average frequencies of variables and the average frequencies between variables, take the parent node as the tail node and the child node as the head node according to the subgraph structure, obtain the probability distribution between nodes using the frequencies and average frequencies, repeat constructing the probability distribution between nodes according to the subgraph structure, take the obtained subgraph probability distribution as the structural prior parameter, and combine with the sample to learn the Bayesian network structure.

[0083] In some embodiments, a penalty factor is added to the scoring function when constructing the structural prior, so that the prior structure is integrated into the posterior structure.

[0084] Such as Figure 4 As shown, an embodiment of the present invention provides a method for sleep disorder attribution analysis based on a knowledge graph, which is applied to an Internet medical platform and includes:

[0085] S201. Collect medical diagnosis and examination data input by users;

[0086] S202, extracting entities from data through an algorithm model;

[0087] S203, using a graph database to store knowledge and building a knowledge graph based on the experience of doctors and experts;

[0088] S204, extracting relevant subgraphs based on the knowledge graph and input information, and constructing a structural prior;

[0089] S205, learning to establish a Bayesian network model based on training samples;

[0090] S206. Use the Bayesian network model established through learning to perform attribution analysis and form analysis results.

[0091] In an embodiment of the present invention, the Internet medical platform may be for psychotherapy or sleep disorders, etc.

[0092] In some embodiments, the Internet medical platform performs preliminary modeling based on historically collected data, and models big data of different disease types, different populations, and different testing schemes by establishing models.

[0093] In some specific examples, the medical diagnosis examination data includes text data and image data.

[0094] In some specific examples, entity extraction includes relationship extraction between sleep disorders and medical diagnosis examination data, extraction of user behavior events, expert experience entity disambiguation, knowledge fusion and knowledge processing of disease and symptom knowledge graphs.

[0095] In some specific examples, subgraphs are extracted and graph neural network models are used to predict the relationships between nodes in the graph to explore more causal relationships.

[0096] In some specific examples, the method combines expert experience to build a knowledge graph of related business fields, extracts relevant subgraphs based on images, uses graph neural network models on subgraphs to predict the relationship between nodes, explores more causal relationships, constructs a priori distribution of Bayesian network organizations, and combines samples to learn Bayesian networks.

[0097] The present invention provides an embodiment of a sleep disorder attribution analysis method based on a knowledge graph, which is applied to an Internet medical platform, including:

[0098] S301, collecting sleep disorder physical examination data input by the user; the physical examination data includes blood data, blood pressure data, urine routine data, electrocardiogram, B-ultrasound, brain CT and other data;

[0099] S302. Use technologies such as NLP, deep learning, and knowledge graphs to perform entity extraction on the data; entity extraction includes the extraction of the relationship between sleep disorders and physical examination data, the extraction of events such as staying up late, caffeine intake, work pressure, and excessive exercise in user behavior events, the disambiguation of doctor experience entities, and the knowledge fusion and knowledge processing of disease and symptom knowledge graphs.

[0100] S303. Use technologies such as graph databases and RDF resource description frameworks to store knowledge, and construct a knowledge graph in combination with doctor expert experience; the knowledge graph includes an overall data graph of different factors, different physical examination data indicators, and corresponding sleep disorders and mental diseases.

[0101] S304. Extract relevant subgraphs according to the knowledge graph and input information, and construct a structural prior.

[0102] S305. Combine training samples to learn and establish a Bayesian network model.

[0103] S306. Use the learned Bayesian network model to perform attribution analysis to form a sleep disorder analysis result.

[0104] In some embodiments, the structural prior is constructed by counting the frequencies of variables (such as electrocardiogram, B-ultrasound, brain CT data, or event variables such as staying up late, caffeine intake, work pressure, and excessive exercise) and the frequencies between variables in large data samples of different patient users, calculating the average frequencies of variables and the average frequencies between variables, taking the parent node as the tail node and the child node as the head node according to the subgraph structure, obtaining the probability distribution between nodes using the frequencies and average frequencies, repeating the construction of the probability distribution between nodes according to the subgraph structure, taking the obtained subgraph probability distribution as the structural prior parameter, and combining with the samples to learn the Bayesian network structure.

[0105] According to an embodiment of another aspect, a sleep disorder attribution analysis system based on a knowledge graph is further provided, including a server side, a client side, and an Internet medical platform.

[0106] The user submits medical diagnosis and examination data through the client side.

[0107] The Internet medical platform collects the medical diagnosis and examination data input by the user, and the server side performs entity extraction on the data through an algorithm model, stores the knowledge using a graph database, and constructs a knowledge graph in combination with doctor expert experience.

[0108] Extract relevant subgraphs according to the knowledge graph and input information, construct a structural prior, combine training samples, and learn and establish a Bayesian network model.

[0109] Use the learned Bayesian network model to perform attribution analysis to form an analysis result.

[0110] In some embodiments, entity extraction includes extraction of the relationship between sleep disorders and medical diagnostic examination data, extraction of user behavior events, disambiguation of expert experience entities, knowledge fusion and knowledge processing of disease and symptom knowledge graphs.

[0111] According to an embodiment of another aspect, there is also provided a computer-readable storage medium, on which a computer program / instructions are stored, and when the computer program / instructions are executed by a processor, the following steps are implemented:

[0112] Perform entity extraction on the collected data through an algorithm model, store the knowledge using a graph database, and construct a knowledge graph in combination with the experience of business experts;

[0113] Extract relevant subgraphs according to the knowledge graph and input information, construct a structural prior, and combine with training samples to learn and establish a Bayesian network model;

[0114] Perform attribution analysis using the learned Bayesian network model.

[0115] According to an embodiment of another aspect, there is also provided a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the following steps are implemented:

[0116] Perform entity extraction on the collected data through an algorithm model, store the knowledge using a graph database, and construct a knowledge graph in combination with the experience of business experts;

[0117] Extract relevant subgraphs according to the knowledge graph and input information, construct a structural prior, and combine with training samples to learn and establish a Bayesian network model;

[0118] Perform attribution analysis using the learned Bayesian network model.

[0119] According to another aspect of the embodiments of this specification, there is also provided an electronic device, including:

[0120] A processor; and

[0121] A memory configured to store computer-executable instructions, and when the executable instructions are executed, the processor performs the following operations:

[0122] Perform entity extraction on the collected data through an algorithm model, store the knowledge using a graph database, and construct a knowledge graph in combination with the experience of business experts;

[0123] Extract relevant subgraphs according to the knowledge graph and input information, construct a structural prior, and combine with training samples to learn and establish a Bayesian network model;

[0124] Perform attribution analysis using the learned Bayesian network model.

[0125] According to another aspect of the embodiments of the present specification, there is also provided a computer-readable storage medium, on which a computer program / instructions are stored, and when the computer program / instructions are executed by a processor, the following steps are implemented:

[0126] Collect medical diagnostic examination data input by a user, perform entity extraction on the data through an algorithm model, store knowledge using a graph database, and construct a knowledge graph in combination with doctors' expert experience;

[0127] According to the knowledge graph and input information, extract relevant subgraphs, construct a structural prior, and learn to establish a Bayesian network model in combination with training samples;

[0128] Perform attribution analysis using the learned Bayesian network model to form an analysis result.

[0129] According to another aspect of the embodiments of the present specification, there is provided a computer program product, including a computer program / instructions, and when the computer program / instructions are executed by a processor, the following steps are implemented:

[0130] Collect medical diagnostic examination data input by a user, perform entity extraction on the data through an algorithm model, store knowledge using a graph database, and construct a knowledge graph in combination with doctors' expert experience;

[0131] According to the knowledge graph and input information, extract relevant subgraphs, construct a structural prior, and learn to establish a Bayesian network model in combination with training samples;

[0132] Perform attribution analysis using the learned Bayesian network model to form an analysis result.

[0133] According to another aspect of the embodiments of the present specification, there is provided an electronic device, including:

[0134] A processor; and

[0135] A memory configured to store computer-executable instructions, and when the executable instructions are executed, the processor performs the following operations:

[0136] Collect medical diagnostic examination data input by a user, perform entity extraction on the data through an algorithm model, store knowledge using a graph database, and construct a knowledge graph in combination with doctors' expert experience;

[0137] According to the knowledge graph and input information, extract relevant subgraphs, construct a structural prior, and learn to establish a Bayesian network model in combination with training samples;

[0138] Perform attribution analysis using the learned Bayesian network model to form an analysis result.

[0139] The sleep disorder attribution analysis method, system and device based on the knowledge graph of the present invention construct a domain knowledge graph, extract subgraphs using the graph, construct structural prior knowledge, and finally perform attribution analysis using the trained Bayesian network, which can effectively utilize expert experience, reduce the requirement of the model for the sample size, improve the performance of the model, accelerate the model training speed, and combine expert experience with model training to obtain effective analysis results in the case of lack of big data on mental diseases or sleep disorders.

[0140] For the convenience of description, when describing the above device, it is divided into various units according to functions for description. Of course, when implementing the present application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0141] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, system, or computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0142] The present invention is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram can be implemented by computer program instructions, and the combination of the flows and / or blocks in the flowchart and / or block diagram can also be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in one Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0143] The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0144] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means that implement the function specified in one or more of the processes and / or blocks Figure 1 in one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0145] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing steps for implementing the function specified in one or more of the processes and / or blocks Figure 1 in one or more of the processes and / or blocks Figure 1 specified in the block or blocks.

[0146] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0147] Memory may include non-permanent memory in the computer-readable medium, random access memory (RAM) and / or non-volatile memory such as read only memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

[0148] Computer-readable media includes both permanent and non-permanent, removable and non-removable media implemented by any method or technology for storage of information such as computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other memory technologies, compact disc read only memory (CD-ROM), digital versatile discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0149] It should also be noted that the term "comprise", "include" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising an..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising said element.

[0150] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and reference can be made to the relevant part of the method embodiment for the related content.

[0151] The above description is only for the embodiments of the present application and is not intended to limit the present application. For those skilled in the art, various modifications and changes can be made to the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for attributing sleep disorders based on a knowledge graph, which is applied to an Internet medical platform, collects sleep disorder physical examination data input by users, and extracts entities from the data through NLP, deep learning, and knowledge graph technologies; entity extraction includes the extraction of the relationship between sleep disorders and physical examination data, the extraction of user behavior events including staying up late, caffeine intake, work pressure, and excessive exercise events, entity disambiguation of doctor experience, knowledge fusion and knowledge processing of disease and symptom knowledge graphs, uses a graph database and RDF resource description framework technology to store knowledge, and constructs a knowledge graph in combination with doctor expert experience; the knowledge graph includes an overall data graph of different factors, different physical examination data indicators, and corresponding sleep disorders and mental diseases; according to the knowledge graph and the input information, relevant subgraphs are extracted, a structural prior is constructed to statistically calculate the frequencies of inspection data variables or external factor variables in the sample and the frequencies between variables, calculate the average frequencies of the variables and the average frequencies between variables, according to the subgraph structure, use the parent node as the tail node and the child node as the head node, obtain the probability distribution between nodes using the frequencies and average frequencies, repeat constructing the probability distribution between nodes, use the obtained subgraph probability distribution as the structural prior parameter, combine sample learning to obtain the Bayesian network structure, and add a penalty factor to the scoring function when constructing the structural prior to integrate the prior structure into the posterior structure; Combined with training samples, learn to establish a Bayesian network model. During the process of learning the Bayesian network structure, add partial mutual information to accurately measure the linear and non-linear causal relationships between variables; Use the established Bayesian network model for attribution analysis to form an analysis result.

2. The method according to claim 1, wherein the physical examination data includes blood data, blood pressure data, urine routine data, electrocardiogram, B-ultrasound, and brain CT data.

3. The method according to claim 1, wherein the inspection data variables include electrocardiogram, B-ultrasound, and brain CT data.

4. The method according to claim 1, wherein the external factor variables include variables such as staying up late, caffeine intake, work pressure, and excessive exercise.

5. The method according to claim 1, combines expert experience, constructs a knowledge graph in the relevant business field, extracts relevant subgraphs based on pictures, uses a graph neural network model on the subgraph to predict the relationships between nodes, mines more causal relationships, constructs a prior distribution of the Bayesian network structure, and combines sample learning of the Bayesian network.

6. A computer-readable storage medium, on which computer programs / instructions are stored, characterized in that, When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1-5 are implemented.

7. A computer program product, comprising computer programs / instructions, characterized in that, When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1-5 are implemented.

8. An electronic device, comprising: Processor; And A memory configured to store computer-executable instructions, the executable instructions, when executed, cause the processor to perform the following operations: Collect sleep disorder physical examination data input by the user, and perform entity extraction on the data through NLP, deep learning, and knowledge graph technologies; entity extraction includes the extraction of the relationship between sleep disorders and physical examination data, the extraction of user behavior events including staying up late, caffeine intake, work pressure, excessive exercise events, the disambiguation of doctor experience entities, the knowledge fusion and knowledge processing of disease and symptom knowledge graphs, and use graph databases and RDF resource description framework technologies to store knowledge, and construct a knowledge graph in combination with doctor expert experience; the knowledge graph includes an overall data graph of different factors, different physical examination data indicators, and corresponding sleep disorders and mental diseases; according to the knowledge graph and input information, extract relevant subgraphs, construct a structural prior to count the frequencies of inspection data variables or external factor variables in the sample and the frequencies between variables, calculate the average frequencies of variables and the average frequencies between variables, according to the subgraph structure, use the parent node as the tail node and the child node as the head node, use the frequencies and average frequencies to obtain the probability distribution between nodes, repeat to construct the probability distribution between nodes, use the obtained subgraph probability distribution as the structural prior parameter, and combine with the sample learning to obtain the Bayesian network structure, and add a penalty factor to the scoring function in the constructed structural prior to make the prior structure fuse into the posterior structure; Combined with training samples, learn to establish a Bayesian network model. During the process of learning the Bayesian network structure, add partial mutual information to accurately measure the linear and non-linear causal relationships between variables; Use the established Bayesian network model for attribution analysis to form an analysis result.

Images