Method and device for identifying correlation based on digestive internal medicine disease

Abstract

The application provides a method and device for identifying the correlation of digestive diseases, which comprises: constructing a graph data structure of digestive diseases; obtaining the comprehensive feature representation of each disease node in the entire graph data structure based on the information propagation and aggregation process of the graph neural network; obtaining the similarity between the feature vector of the clinical data of the patient to be identified and the disease node feature vector; and obtaining the correlation degree between the patient to be identified and different diseases based on the similarity value. The application accurately evaluates the potential correlation degree between the patient and different digestive diseases, provides an objective and quantitative basis for the doctor's decision, and assists the doctor in more scientifically judging the disease that the patient is most likely to have when facing complex clinical symptoms and multiple possible diseases, thereby reducing the probability of misdiagnosis and missed diagnosis.

Description

Technical Field

[0001] This application relates to the field of association identification technology, specifically a method and device for association identification based on digestive diseases. Background Technology

[0002] In the field of gastroenterology, identifying associations among gastrointestinal diseases has always been a significant challenge for healthcare professionals. Traditional methods rely primarily on physicians' personal experience and preliminary assessments of patient symptoms and medical history, combined with limited diagnostic tools. However, this approach has several limitations. Firstly, gastrointestinal symptoms are complex, diverse, and often nonspecific; many diseases present with similar early symptoms, leading to misdiagnosis and missed diagnosis. For example, abdominal pain, a common symptom, may be associated with multiple diseases such as gastric ulcers, cholecystitis, appendicitis, and intestinal inflammation, making it difficult to accurately determine the specific cause based on a single symptom. Secondly, with the continuous development of medical technology, various diagnostic methods generate massive amounts of clinical data. However, this data is often scattered across different hospital information systems, lacking effective integration and analysis methods. Physicians struggle to comprehensively and accurately acquire and utilize this data to aid diagnosis within a short period. Therefore, developing a solution that can effectively integrate knowledge and data and accurately identify associations among gastrointestinal diseases is of significant practical importance and an urgent clinical need.

[0003] Therefore, overcoming the aforementioned technical problems and defects has become a key issue that needs to be addressed. Summary of the Invention

[0004] To overcome the aforementioned problems in the prior art, this application provides a method and apparatus for association identification based on digestive diseases, employing the following technical solution:

[0005] Firstly, this application provides a method for identifying associations based on gastrointestinal diseases, including:

[0006] Constructing a graph data structure for gastroenterological diseases;

[0007] Based on the information propagation and aggregation process of graph neural networks, a comprehensive feature representation of each disease node in the entire graph data structure is obtained;

[0008] The similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node is obtained, and the degree of association between the patient to be identified and different diseases is obtained based on the similarity value.

[0009] Furthermore, the construction of the graph data structure for gastroenterological diseases includes:

[0010] Data related to gastroenterological diseases is obtained from different channels, and entities in the data are identified, including disease entities, symptom entities, examination entities, and treatment entities.

[0011] Relationships between entities are obtained based on clustering algorithms;

[0012] The identified entities are treated as nodes, and the relationships between entities are treated as edges, thus constructing a graph data structure.

[0013] Furthermore, the specific content of obtaining the relationships between entities based on the clustering algorithm includes:

[0014] Define an entity feature vector based on the features that describe the properties of each entity.

[0015] Standardize the entity feature vectors to obtain standardized features;

[0016] Core point determination: Select any unvisited entity point P from the dataset, calculate the distance from P to all other entity points. If there are at least MinPts entity points (including P itself) within the neighborhood radius ε of P, then P is marked as a core point.

[0017] Cluster expansion: With the core point P as the center, find all points in the neighborhood of the core point P to form a temporary cluster; check whether each point in the temporary cluster is a core point, and add the points in its neighborhood that meet the condition to the temporary cluster; continue recursively until there are no new points that can be added to the temporary cluster.

[0018] Noise point labeling: Points that are neither core points nor included in any clusters are labeled as noise points;

[0019] Continue to select the next unvisited point and repeat the process of core point determination, cluster expansion and noise point marking until all entity points have been visited and obtain the entity clustering results.

[0020] Analyze the common characteristics of entities within each cluster to determine the close relationships between entities.

[0021] Furthermore, the information propagation and aggregation process based on graph neural networks, which obtains a comprehensive feature representation of each disease node in the entire graph data structure, includes:

[0022] Calculate the attention coefficient: Using the graph data structure as input to the graph neural network, in the first layer of the graph neural network, for each disease node i and its neighboring nodes j, their feature vectors h are calculated. i and h j Perform concatenation to obtain a new vector [h] i ||h j ], through the preset line aT ([h i ||h j Attention coefficient e is obtained through gender transformation and preset activation function. ij The attention coefficient e ij This indicates the importance of neighbor node j to disease node i.

[0023] Furthermore, the information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes:

[0024] Normalization: The attention coefficient e of all neighboring nodes j for each disease node i ij Normalization is performed to obtain the attention weights. in Ni Let i be the set of neighboring nodes of node i. The attention weight of each neighboring node is between 0 and 1, and the sum of the attention weights of all neighboring nodes of the same disease node is 1.

[0025] Furthermore, the information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes:

[0026] Information aggregation and node updating: based on the attention mechanism α ij We perform weighted aggregation of the features of neighboring nodes and update the feature representation of the disease node as follows: Where W is a learnable weight matrix and σ is an activation function (such as ReLU). In this way, disease nodes can selectively aggregate information based on the importance of their neighboring nodes, highlighting key associations.

[0027] Furthermore, the information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes:

[0028] Using the node feature vector output from the previous layer as the input to the next layer, the steps of calculating attention coefficients, normalization, information aggregation, and node updating are repeated. As the number of layers increases, disease nodes can incorporate more information from skip neighbors.

[0029] After the information is propagated and aggregated through all layers of the graph neural network, the disease node feature vector output by the last layer is the comprehensive feature representation in the entire graph data structure.

[0030] Secondly, this application also provides a device for identifying associations based on digestive diseases, comprising:

[0031] The graph data structure construction module is used to build graph data structures for digestive diseases.

[0032] The comprehensive feature representation acquisition module is used to acquire the comprehensive feature representation of each disease node in the entire graph data structure based on the information propagation and aggregation process of graph neural networks.

[0033] The correlation degree acquisition module is used to obtain the similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node, and to obtain the correlation degree between the patient to be identified and different diseases based on the similarity value.

[0034] Thirdly, this application provides an electronic device, comprising:

[0035] One or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory, and the one or more computer programs include instructions that, when executed by the device, cause the device to perform the method as described in the first aspect.

[0036] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when run on a computer, causes the computer to perform the method described in the first aspect.

[0037] Fifthly, this application provides a computer program that, when executed by a computer, performs the method described in the first aspect.

[0038] In one possible design, the program in the fifth aspect can be stored wholly or partially on a storage medium packaged with the processor, or it can be stored wholly or partially on a memory not packaged with the processor.

[0039] This application has the following beneficial effects:

[0040] 1. This application constructs a graph data structure for gastroenterological diseases, visually representing the complex relationships between various gastroenterological conditions, such as etiological associations, symptom similarities, and complication relationships. Compared to traditional tables or text formats, graph data structures can more clearly present the interrelationships between diseases, facilitating physicians and researchers to grasp the overall disease relationship network. This application integrates information from different data sources, such as professional medical literature, clinical guidelines, hospital electronic medical records, and expert experience, transforming these multi-dimensional and heterogeneous data into nodes and edges in a graph data structure, achieving unified data representation and management.

[0041] 2. This application obtains the comprehensive feature representation of each disease node in the entire graph data structure through the information propagation and aggregation process based on graph neural networks. This application automatically learns the mutual influence and correlation between disease nodes through graph neural networks, and makes full use of the global structural information of the graph to generate the feature representation of each disease node, thereby describing the characteristics of the disease more comprehensively and accurately. In gastroenterology diseases, multimodal data such as patients' gastroscopy images, pathology report texts, and physiological index values ​​are combined and fused and learned through graph neural networks to obtain richer and more comprehensive disease node feature representations.

[0042] 3. This application obtains the similarity between the feature vector of the patient's clinical data and the feature vector of disease nodes, and uses the similarity value to determine the degree of association between the patient and different diseases. This provides personalized disease association analysis for the patient. By calculating the similarity between the patient's clinical data features and disease node features, the potential association between the patient and different gastroenterological diseases can be accurately assessed, providing objective and quantitative evidence for physicians' decision-making. This application intuitively reflects the degree of matching between the patient and various diseases through similarity values, assisting physicians in more scientifically determining the most likely disease when faced with complex clinical symptoms and multiple possible disease diagnoses, reducing the probability of misdiagnosis and missed diagnosis, and improving the accuracy and reliability of diagnosis. Attached Figure Description

[0043] Figure 1 This is an exemplary system architecture diagram to which embodiments of this application can be applied;

[0044] Figure 2 This is a flowchart of the association identification method based on gastroenterological diseases according to an embodiment of this application;

[0045] Figure 3 This is a flowchart illustrating the construction graph data structure of an embodiment of this application;

[0046] Figure 4 This is a flowchart illustrating the process of obtaining entity relationships in an embodiment of this application.

[0047] Figure 5 This is a flowchart of the apparatus according to an embodiment of this application;

[0048] Figure 6 This is a schematic diagram of a computer device according to an embodiment of this application. Detailed Implementation

[0049] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0050] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0051] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0052] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0053] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0054] Terminal devices 101, 102, and 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc.

[0055] Server 105 can be a server that provides various services, such as a backend server that supports the pages displayed on terminal devices 101, 102, and 103.

[0056] It should be noted that the association identification method based on gastroenterological diseases provided in this application embodiment is generally executed by a server / terminal device, and correspondingly, the association identification device based on gastroenterological diseases is generally set in the server / terminal device.

[0057] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0058] Continue to refer to Figure 2 The figure shows a flowchart of a method for identifying associations based on gastrointestinal diseases according to this application. The method includes the following steps:

[0059] Step 201: Construct a graph data structure for gastroenterological diseases.

[0060] In one possible implementation, the construction of the graph data structure for gastroenterological diseases is described in reference [reference needed]. Figure 3 The specific content includes:

[0061] Step 31: Obtain relevant data on gastroenterological diseases from different channels and identify entities in the relevant data, including disease entities, symptom entities, examination entities, and treatment entities.

[0062] For example, different channels include: professional medical literature, clinical guidelines, hospital electronic medical records, and expert experience. Among them, medical literature provides theoretical knowledge and research results on diseases, clinical guidelines contain standardized diagnosis and treatment procedures and recommendations, electronic medical records record the clinical information of real patients, and expert experience can supplement some special cases and complex cases.

[0063] In the process of identifying entities in relevant data, including disease entities, symptom entities, examination entities, and treatment entities, entities can be extracted by Named Entity Recognition (NER), or identified based on rules (e.g., matching using a medical terminology dictionary), machine learning methods (such as Conditional Random Field (CRF) models), or deep learning models (such as Bi-LSTM bidirectional long short-term memory networks).

[0064] Step 32: Obtain the relationships between entities based on clustering algorithms.

[0065] In the embodiments of this application, for example, the relationships between entities include "disease-symptom" (e.g., "gastric ulcer - manifested as - abdominal pain"), "disease-examination" (e.g., "gastric ulcer - diagnosed by - gastroscopy"), "disease-treatment" (e.g., "gastric ulcer - treated with - proton pump inhibitor"), "symptom-examination" (e.g., "abdominal pain - to find the cause by - abdominal ultrasound"), etc.

[0066] In one possible implementation, the step of obtaining the relationships between entities based on a clustering algorithm is described in reference [link to relevant documentation]. Figure 4 The specific content includes:

[0067] Step 41: Define entity feature vectors based on the features that describe the properties of each entity.

[0068] For example, a disease entity may include characteristics such as the location of the disease (e.g., esophagus, stomach, intestines), the type of disease (e.g., inflammation, tumor, functional), and the severity (mild, moderate, severe); a symptom entity may have characteristics such as the location of the symptom, the nature of the symptom (e.g., pain, bloating, nausea), and the frequency of attacks; an examination entity may have characteristics such as the examination method (e.g., imaging examination, laboratory examination), the accuracy of the examination, and the scope of application; and a treatment entity may include characteristics such as the type of treatment (e.g., drug treatment, surgical treatment), the treatment effect, and side effects.

[0069] Step 42: Standardize the entity feature vector to obtain standardized features.

[0070] In this embodiment, the entity feature vectors are standardized to make the numerical ranges of different features comparable. If one feature is the incidence rate (ranging from 0 to 100%) and another feature is the examination accuracy (ranging from 0 to 1), they can both be normalized to the 0-1 range. For categorical features (such as disease types like inflammation and tumors), one-hot encoding can be used to convert them into numerical form.

[0071] Step 43, Core Point Determination: Select any unvisited entity point P from the dataset, calculate the distance from P to all other entity points. If there are at least MinPts entity points (including P itself) within the neighborhood radius ε of P, then P is marked as a core point.

[0072] Step 44, Cluster Expansion: With the core point P as the center, find all points in the neighborhood of the core point P to form a temporary cluster; check whether each point in the temporary cluster is a core point, and add the points in its neighborhood that meet the conditions to the temporary cluster; continue recursively until there are no new points that can be added to the temporary cluster.

[0073] Step 45, Noise Point Labeling: Points that are neither core points nor included in any cluster are labeled as noise points. These points may be outliers or isolated entities in the data; in the gastroenterology knowledge graph, they may represent rare disease symptoms or specialized examination methods.

[0074] Step 46: Continue to select the next unvisited point and repeat the above process of core point determination, cluster expansion and noise point marking until all entity points have been visited and obtain the entity clustering results.

[0075] Step 47: Analyze the common characteristics of entities within each cluster to determine the close relationships between entities. For example, consider a cluster containing multiple gastrointestinal diseases. These diseases may share similar symptoms, diagnostic methods, or treatments. For instance, diseases within a cluster may all be related to abdominal pain and require similar blood tests for diagnosis. This suggests that these diseases share similarities in clinical presentation and diagnostic methods, and may have common pathogenesis or pathophysiological processes.

[0076] In one possible implementation, common characteristics of entities within each cluster are analyzed to determine close relationships between entities, including entity relationship definition and entity relationship analysis.

[0077] The definition of entity relationships includes: defining relationships between entities based on clustering results. When two entities are clustered in the same cluster, they are considered to have similar features, and a similarity relationship is defined between them. For example, if the clustering results of a disease entity frequently coincide with the clustering results of certain symptom entities, an association relationship is defined between the disease entity and these symptom entities; similarly, if the clustering results of a treatment entity correspond to the clustering results of certain disease entities, a treatment relationship is defined between the treatment entity and the disease entity.

[0078] Entity relationship analysis includes analyzing and interpreting the extracted entity relationships. Visualization tools, such as relationship diagrams and heatmaps, are used to visually represent the relationships between entities, helping doctors and researchers understand the intrinsic connections between entities related to gastroenterological diseases. For example, analyzing the relationship between disease entities and symptom entities can reveal typical symptom combinations for certain diseases; analyzing the relationship between disease entities and treatment entities can provide insights into commonly used treatment methods and their effects for different diseases.

[0079] Step 33: Construct a graph data structure by using the identified entities as nodes and the relationships between entities as edges.

[0080] This application presents the complex relationships between various diseases in gastroenterology, such as etiological associations, symptom similarities, and complication relationships, using an intuitive graph structure. Compared to traditional tables or text formats, graph data structures can more clearly present the interrelationships between diseases, making it easier for doctors and researchers to grasp the overall disease relationship network. This application integrates information from different data sources, such as professional medical literature, clinical guidelines, hospital electronic medical records, and expert experience, transforming these multi-dimensional and heterogeneous data into nodes and edges in a graph data structure, achieving unified data representation and management.

[0081] Step 202: Based on the information propagation and aggregation process of the graph neural network, obtain the comprehensive feature representation of each disease node in the entire graph data structure.

[0082] In one possible implementation, the information propagation and aggregation process based on graph neural networks, which obtains a comprehensive feature representation of each disease node within the entire graph data structure, includes:

[0083] Calculate the attention coefficient: Using the graph data structure as input to the graph neural network, in the first layer of the graph neural network, for each disease node i and its neighboring nodes j, their feature vectors h are calculated. i and h j Perform concatenation to obtain a new vector [h] i ||h j ], through the preset line a T ([h i ||h j Attention coefficient e is obtained through gender transformation and preset activation function. ij The attention coefficient e ij This indicates the importance of neighbor node j to disease node i. For example, in a gastroenterology disease knowledge graph, the attention coefficient between the disease node "gastric ulcer" and the symptom node "abdominal pain" may be high because abdominal pain is a typical symptom of gastric ulcer.

[0084] Normalization: The attention coefficient e of all neighboring nodes j for each disease node i ij Normalization is performed to obtain the attention weights. in Ni Let i be the set of neighboring nodes of node i; thus, the attention weight of each neighboring node is between 0 and 1, and the sum of the attention weights of all neighboring nodes of the same disease node is 1.

[0085] Information aggregation and node updating: based on the attention mechanism α ij We perform weighted aggregation of the features of neighboring nodes and update the feature representation of the disease node as follows: Here, W is a learnable weight matrix, and σ is an activation function (such as ReLU). In this way, disease nodes can selectively aggregate information based on the importance of their neighboring nodes, highlighting key associations. For example, a "gastric ulcer" disease node will pay more attention to information from closely related symptom nodes such as "abdominal pain" and "acid reflux," and these key pieces of information will have higher weights in the updated feature representation.

[0086] In one possible implementation, the information propagation and aggregation process based on graph neural networks, which obtains a comprehensive feature representation of each disease node in the entire graph data structure, further includes:

[0087] Using the node feature vectors output from the previous layer as input to the next layer, the steps of calculating attention coefficients, normalization, information aggregation, and node updates are repeated. As the number of layers increases, disease nodes can incorporate more information from skip neighbors. For example, in the second layer, disease nodes can not only aggregate directly related symptom node information, but also aggregate information from other connected examination or treatment nodes through these symptom nodes, thereby obtaining more comprehensive correlation information.

[0088] After information is propagated and aggregated through all layers of the graph neural network, the disease node feature vector output by the last layer is the comprehensive feature representation in the entire graph data structure. This comprehensive feature representation integrates the initial features of the disease node itself with information obtained from neighboring nodes (symptom, examination, and treatment nodes) in the graph, comprehensively reflecting the disease's position and associations within the knowledge graph. For example, the comprehensive feature representation of the "gastric ulcer" disease node might include its association strength with typical symptoms (abdominal pain, acid reflux), the correlation with diagnostic examinations (gastroscopy), and information on commonly used treatment methods (drug therapy). This information is encoded in the feature vector for subsequent disease association identification.

[0089] This application uses graph neural networks to automatically learn the mutual influence and correlation between disease nodes, making full use of the global structural information of the graph to generate feature representations of each disease node, thereby describing the characteristics of the disease more comprehensively and accurately. In gastroenterology diseases, multimodal data such as patients' gastroscopy images, pathology report texts, and physiological index values ​​are combined and fused and learned through graph neural networks to obtain richer and more comprehensive feature representations of disease nodes.

[0090] Step 203: Obtain the similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node, and obtain the degree of association between the patient to be identified and different diseases based on the similarity value.

[0091] In this embodiment of the application, the patient's clinical data includes the patient's medical records, examination reports, etc. The patient's clinical data covers symptoms (such as abdominal pain, nausea, vomiting, etc.), examination results (such as gastroscopy, blood test indicators, etc.), and medical history information (such as past medical history, family medical history, etc.). For example, information collected includes that the patient has "abdominal pain" symptoms lasting "3 days", gastroscopy shows "gastric mucosal erosion", and blood tests show "elevated white blood cell count", etc.

[0092] In one possible implementation, the step of obtaining the similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node, and obtaining the degree of association between the patient to be identified and different diseases based on the similarity value, specifically includes:

[0093] The similarity between the patient's feature vector and the feature vectors of all disease nodes is calculated based on a preset formula, and then sorted from high to low similarity values. The similarity calculation formula is as follows: in These are the patient feature vector and the disease node feature vector, respectively. The cosine similarity value ranges from -1 to 1; the closer the value is to 1, the more similar the two vectors are.

[0094] The calculated similarity scores between the patients and each disease node are then ranked. For example, if the cosine similarity between a patient and the "gastric ulcer" disease node is 0.8, the cosine similarity with the "gastritis" disease node is 0.6, and the cosine similarity with the "cholecystitis" disease node is 0.4, then the patients are ranked from highest to lowest similarity as "gastric ulcer" > "gastritis" > "cholecystitis".

[0095] Based on the similarity ranking, diseases ranked higher are more closely associated with patients, while those ranked lower are less closely associated. For example, in the above case, the patient is most closely associated with "gastric ulcer," suggesting a relatively high probability that the patient has a "gastric ulcer." The disease ranked first has the strongest association with the patient and can be considered the primary diagnostic direction; diseases ranked lower have a weaker association, but cannot be completely ruled out.

[0096] This application calculates the similarity between a patient's clinical data characteristics and disease node characteristics, accurately assessing the potential association between a patient and various gastroenterological diseases, providing objective and quantitative evidence for physicians' decision-making. The similarity value intuitively reflects the degree of matching between a patient and various diseases, assisting physicians in more scientifically determining the most likely disease a patient has when faced with complex clinical symptoms and multiple possible disease scenarios, reducing the probability of misdiagnosis and missed diagnosis, and improving the accuracy and reliability of diagnosis.

[0097] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0098] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0099] Continue to refer to Figure 5 The gastroenterology-based disease association identification device described in this embodiment includes:

[0100] Graph data structure construction module 501 is used to construct graph data structures for digestive diseases.

[0101] The comprehensive feature representation acquisition module 502 is used for the information propagation and aggregation process based on graph neural networks to acquire the comprehensive feature representation of each disease node in the entire graph data structure.

[0102] The correlation degree acquisition module 503 is used to obtain the similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node, and to obtain the correlation degree between the patient to be identified and different diseases based on the similarity value.

[0103] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed] for details. Figure 6 , Figure 6 This is a basic structural block diagram of the computer device in this embodiment.

[0104] The computer device 6 includes a memory 6a, a processor 6b, and a network interface 6c that are interconnected via a system bus. It should be noted that only the computer device 6 with components 6a-6c is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0105] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0106] The memory 6a includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 6a may be an internal storage unit of the computer device 6, such as the hard disk or memory of the computer device 6. In other embodiments, the memory 6a may also be an external storage device of the computer device 6, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 6. Of course, the memory 6a may also include both the internal storage unit and its external storage device of the computer device 6. In this embodiment, the memory 6a is typically used to store the operating system and various application software installed on the computer device 6, such as program code for a method for identifying associations of digestive diseases. In addition, the memory 6a can also be used to temporarily store various types of data that have been output or will be output.

[0107] In some embodiments, the processor 6b may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. The processor 6b is typically used to control the overall operation of the computer device 6. In this embodiment, the processor 6b is used to run program code stored in the memory 6a or process data, for example, to run the program code for the association identification method based on digestive diseases.

[0108] The network interface 6c may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 6 and other electronic devices.

[0109] This application also provides another embodiment, namely, a non-volatile computer-readable storage medium storing a program for a method for identifying associations based on gastrointestinal diseases, wherein the method for identifying associations based on gastrointestinal diseases can be executed by at least one processor to perform the steps of the method for identifying associations based on gastrointestinal diseases as described above.

[0110] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0111] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. A method for identifying associations based on gastrointestinal diseases, characterized in that, include: Constructing a graph data structure for gastroenterological diseases; Based on the information propagation and aggregation process of graph neural networks, a comprehensive feature representation of each disease node in the entire graph data structure is obtained; The similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node is obtained, and the degree of association between the patient to be identified and different diseases is obtained based on the similarity value.

2. The method for association identification based on gastroenterological diseases according to claim 1, characterized in that, The construction of the graph data structure for digestive diseases includes: Data related to gastroenterological diseases is obtained from different channels, and entities in the data are identified, including disease entities, symptom entities, examination entities, and treatment entities. Relationships between entities are obtained based on clustering algorithms; The identified entities are treated as nodes, and the relationships between entities are treated as edges, thus constructing a graph data structure.

3. The method for association identification based on gastroenterological diseases according to claim 2, characterized in that, The method for obtaining relationships between entities based on clustering algorithms specifically includes: Define an entity feature vector based on the features that describe the properties of each entity. Standardize the entity feature vectors to obtain standardized features; Core point determination: Select any unvisited entity point P from the dataset, calculate the distance from P to all other entity points. If there are at least MinPts entity points (including P itself) within the neighborhood radius ε of P, then P is marked as a core point. Cluster expansion: With the core point P as the center, find all points in the neighborhood of the core point P to form a temporary cluster; check whether each point in the temporary cluster is a core point, and add the points in its neighborhood that meet the condition to the temporary cluster; continue recursively until there are no new points that can be added to the temporary cluster. Noise point labeling: Points that are neither core points nor included in any clusters are labeled as noise points; Continue to select the next unvisited point and repeat the process of core point determination, cluster expansion and noise point marking until all entity points have been visited and obtain the entity clustering results. Analyze the common characteristics of entities within each cluster to determine the close relationships between entities.

4. The method for association identification based on gastroenterological diseases according to claim 1, characterized in that, The information propagation and aggregation process based on graph neural networks obtains a comprehensive feature representation of each disease node within the entire graph data structure, including: Calculate the attention coefficient: Using the graph data structure as input to the graph neural network, in the first layer of the graph neural network, for each disease node i and its neighboring nodes j, their feature vectors h are calculated. i and h j Perform concatenation to obtain a new vector [h] i ||h j ], through the preset line a T ([h i ||h j Attention coefficient e is obtained through gender transformation and preset activation function. ij The attention coefficient e ij This indicates the importance of neighbor node j to disease node i.

5. The method for association identification based on gastroenterological diseases according to claim 4, characterized in that, The information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes: Normalization: The attention coefficient e of all neighboring nodes j for each disease node i ij Normalization is performed to obtain the attention weights. in Ni Let i be the set of neighboring nodes of node i. The attention weight of each neighboring node is between 0 and 1, and the sum of the attention weights of all neighboring nodes of the same disease node is 1.

6. The method for association identification based on gastroenterological diseases according to claim 5, characterized in that, The information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes: Information aggregation and node updating: based on the attention mechanism α ij We perform weighted aggregation of the features of neighboring nodes and update the feature representation of the disease node as follows: Where W is a learnable weight matrix and σ is the activation function.

7. The method for association identification based on gastroenterological diseases according to claim 6, characterized in that, The information propagation and aggregation process based on graph neural networks, which obtains the comprehensive feature representation of each disease node in the entire graph data structure, also includes: Using the node feature vector output from the previous layer as the input to the next layer, the steps of calculating attention coefficients, normalization, information aggregation, and node updating are repeated. As the number of layers increases, disease nodes can incorporate more information from skip neighbors. After the information is propagated and aggregated through all layers of the graph neural network, the disease node feature vector output by the last layer is the comprehensive feature representation in the entire graph data structure.

8. A device for identifying associations based on digestive diseases, used to implement the methods for identifying associations based on digestive diseases according to claims 1-7, characterized in that, include: The graph data structure construction module is used to construct graph data structures for digestive diseases. The comprehensive feature representation acquisition module is used to acquire the comprehensive feature representation of each disease node in the entire graph data structure based on the information propagation and aggregation process of graph neural networks. The correlation degree acquisition module is used to obtain the similarity between the feature vector of the clinical data of the patient to be identified and the feature vector of the disease node, and to obtain the correlation degree between the patient to be identified and different diseases based on the similarity value.

9. An electronic device, characterized in that, include: One or more processors; Memory; And one or more computer programs, wherein the one or more computer programs are stored in the memory, the one or more computer programs including instructions that, when executed by the device, cause the device to perform the steps of the method for associating gastrointestinal diseases as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when run on a computer, causes the computer to perform the steps of the association identification method based on digestive diseases as described in any one of claims 1 to 7.

Images