Traditional Chinese medicine pre-inquiry collection management system

Abstract

The invention is suitable for the technical field of medical treatment, and provides a traditional Chinese medicine pre-inquiry collection management system, which comprises a user information acquisition module, a traditional Chinese medicine constitution identification module, a symptom association analysis module, a personalized inquiry path generation module, a data integration and medical record generation module and a feedback optimization module. A constitution identification model is constructed by collecting basic information and chief complaint symptoms of a user, and the constitution type is accurately identified; a high-correlation symptom sequence is generated based on a knowledge graph and multi-dimensional weight calculation, an interactive inquiry tree is constructed through a dynamic branch algorithm, and dynamic adjustment of a personalized inquiry path is achieved; real-time data and historical diagnosis and treatment records are integrated to generate a structured pre-diagnosis medical record, and the knowledge graph weight is fed back and optimized through clinical definite diagnosis data, so that the diagnosis precision is continuously improved. Through data fusion and an intelligent algorithm, the objectivity and accuracy of traditional Chinese medicine diagnosis are remarkably improved, missed diagnosis and misdiagnosis are reduced, and scientific and personalized auxiliary decision support is provided for doctors.

Description

Technical Field

[0001] The invention belongs to the field of medical technology, and in particular relates to a traditional Chinese medicine pre-consultation collection and management system. Background Art

[0002] With the rapid development of information technology and artificial intelligence, the application of intelligent medical technology in TCM diagnosis and treatment has gradually become a research hotspot. Traditional TCM diagnosis relies on the doctor's experience and subjective judgment of the patient's symptoms. However, the subjectivity and experience-based reliance in the diagnostic process make the diagnostic results susceptible to individual differences, situational factors and doctor experience.

[0003] However, in the existing technology, most intelligent medical consultation systems still face problems such as inaccurate symptom recognition, inaccurate identification of physical constitution, and insufficient analysis of the correlation between symptoms and physical constitution. In addition, existing systems often lack dynamic optimization mechanisms, making it difficult to make feedback adjustments based on the differences between clinical diagnosis data and pre-diagnosis data, which results in limited diagnostic accuracy of the system and cannot effectively adapt to the diverse needs of different individuals. Summary of the invention

[0004] The purpose of the present invention is to provide a TCM pre-consultation collection and management system, aiming to solve the technical problems existing in the prior art identified in the background technology.

[0005] A TCM pre-consultation collection and management system, comprising a user information collection module, a TCM constitution identification module, a symptom association analysis module, a personalized consultation path generation module, a data integration and medical record generation module and a feedback optimization module, wherein; The user information collection module is used to obtain basic information and main symptoms of the user and generate an initial data set including a symptom feature vector; The TCM constitution identification module receives the symptom feature vector in the initial data set, combines the tongue image and pulse parameters, and outputs the constitution type and the associated symptom group; The symptom association analysis module calls the corresponding knowledge graph according to the constitution type, performs multi-dimensional weight calculation on the associated symptom group, and generates a high-association symptom sequence; The personalized consultation path generation module generates an interactive consultation tree including mandatory symptom verification items and optional symptom extension items based on the highly correlated symptom sequence through a dynamic branching algorithm; The data integration and medical record generation module integrates the interactive data of the medical inquiry tree and the historical medical records to generate a pre-diagnosis medical record including a physical characteristic marker and a symptom evolution path; The feedback optimization module is used to obtain the difference characteristics between clinical diagnosis data and pre-diagnosis medical records, and adjust the knowledge graph weight coefficient through the parameter self-correction unit.

[0006] The beneficial effects of the present invention are: The present invention can accurately identify and match the user's constitution type and related symptoms through the collaborative work of the user information collection module, the traditional Chinese medicine constitution identification module and the symptom association analysis module. In particular, in the constitution identification process, the user's basic information, main symptoms, tongue and pulse parameters are combined, and a multi-integrated constitution identification model is constructed to ensure efficient matching of constitution types (such as yin deficiency, yang deficiency and qi deficiency) and symptom groups. Through the high-correlation symptom sequence and personalized consultation path generated by the system, the user's diagnosis can be made more accurate, reducing the probability of misdiagnosis or missed diagnosis.

[0007] The present invention realizes a dynamic optimization function based on the difference between clinical diagnosis data and pre-diagnosis medical records through a feedback optimization module. By adjusting the weight coefficient of the knowledge graph, the optimization module can provide real-time feedback on the correlation between the user's symptom manifestation and constitution type, and perform self-correction according to the actual diagnosis and treatment results. This mechanism can effectively improve the diagnostic accuracy of the system and ensure that as more user data is accumulated, the system's constitution identification and symptom association analysis functions are continuously optimized, thereby providing doctors with more accurate support and optimizing the diagnosis and treatment process.

[0008] Through the data integration and medical record generation module, the system of the present invention can effectively integrate the user's inquiry tree interaction data with historical diagnosis and treatment records to generate a pre-diagnosis medical record including physical characteristic markers and symptom evolution paths. This medical record can not only reflect the user's current symptoms and physical characteristics, but also provide a complete symptom evolution process in combination with historical records, providing doctors with detailed diagnosis and treatment basis. The generation of this medical record not only improves the efficiency of diagnosis, but also provides complete reference data for subsequent treatment and follow-up visits, further improving the quality and accuracy of medical services. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A schematic diagram of a flow chart of a TCM pre-consultation collection and management system provided by an embodiment of the present invention; Figure 2 This is a schematic diagram of the flow of the TCM constitution identification module provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0010] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0011] like Figure 1-Figure 2As shown, a TCM pre-consultation collection and management system includes a user information collection module, a TCM constitution identification module, a symptom association analysis module, a personalized consultation path generation module, a data integration and medical record generation module and a feedback optimization module, wherein the operation process of the TCM pre-consultation collection and management system is as follows: User information collection: Users enter basic information (age, gender, lifestyle, etc.) and main symptoms (such as pain location, frequency of attacks, etc.) through an interactive interface (such as mobile, web or hospital terminal equipment). The system converts the information into a numerical symptom feature vector, generates an initial data set and stores it in the database.

[0012] TCM constitution identification: The system calls the symptom feature vector in the initial data set, combines the tongue image parameters (obtaining the tongue quality, tongue coating, and tongue shape through the tongue image acquisition device) and pulse parameters (obtaining the speed, strength, and fluctuation shape of the pulse through the pulse diagnosis device), and uses the support vector machine algorithm (SVM) to build a constitution identification model to output the user's constitution type (such as Yin deficiency, Yang deficiency, Qi deficiency) and related symptom groups.

[0013] Symptom association analysis: According to the constitution type, the corresponding TCM symptom knowledge graph is called to perform multi-dimensional weight calculations on the associated symptoms (including the correlation between symptoms and constitution, the similarity between symptoms, and the weight of the joint relationship), and a high-correlation symptom sequence sorted by weight is generated (the first 20% are mandatory verification items, and the remaining first 20% are optional extension items).

[0014] Personalized consultation path generation: An interactive consultation tree is constructed through a dynamic branching algorithm, with mandatory symptom verification items as core nodes and optional symptom extension items as sub-branches. The system dynamically adjusts the consultation path based on real-time user feedback (such as symptom changes, doctor input) to generate a multi-level, scalable personalized process.

[0015] Data integration and medical record generation: Integrate real-time consultation data with the user's historical medical records (such as previous physical type and symptom evolution history) to generate a structured pre-diagnosis medical record containing physical characteristic markers, symptom evolution path and treatment recommendations, and output it to the doctor's terminal.

[0016] Feedback optimization: Compare the difference characteristics between clinical diagnosis data and pre-diagnosis medical records (such as symptom correlation deviation and physical constitution misjudgment), adjust the knowledge graph weight coefficient through the parameter self-correction unit, and optimize the subsequent diagnosis logic.

[0017] Specific: User information collection module

[0018] Obtaining basic user information: Obtaining basic user information through user input devices or interactive interfaces, including age, gender, height, weight, living habits (such as diet, sleep, etc.), family medical history, etc.; Obtaining the main symptoms: Collect the user's main symptoms through questionnaires, interviews or form input. The symptoms include detailed information such as the body part of the discomfort, pain level, onset duration, symptom duration, and symptom frequency; Converting into symptom feature vector: The basic information and main complaint symptom information input by the user are converted into symptom feature vector through the data processing unit. The symptom feature vector includes the user's physical information (such as age, gender, etc.), main complaint symptom information (such as symptom type, degree, time sequence, etc.) and symptom-related background information (such as the time period of symptom occurrence, inducement, etc.); Generate an initial data set: Integrate the generated symptom feature vector with the user's basic information to construct an initial data set, which includes each user's symptom feature vector and its basic information as input data for subsequent processing; Upload and store data: upload the generated symptom feature vector and user's basic information to the system database for storage; The user information collection module collects the user's basic information and main symptoms through systematic steps and converts them into symptom feature vectors, ensuring the accuracy and completeness of the information, and providing a solid data foundation for subsequent constitution identification, symptom analysis, and the generation of personalized consultation paths.

[0019] Data input methods: Support multiple input methods, including electronic forms, voice input (such as converting interview recordings to text), and sensor devices (such as wearable devices to collect sleep data).

[0020] Symptom feature vector generation: The main symptom text (such as "dry mouth") is converted into a numerical vector through natural language processing (NLP) technology, for example: Physical information vector: age (numeric), gender (0 / 1 coding), lifestyle habits (diet score, sleep duration).

[0021] Symptom vector: pain intensity (1-10 points), duration of attack (hours), and associated context (such as "worsened at night" coded as a feature value).

[0022] TCM constitution identification module Receiving a symptom feature vector: obtaining an initial data set from a user information collection module, which includes the user's symptom feature vector; Obtain tongue parameters: Obtain tongue parameters through tongue image collection equipment or manual inspection. Tongue parameters include tongue quality (such as red, light, tender, etc.), tongue coating (such as white, yellow, thin and thick, etc.) and tongue shape (such as tongue size, cracks on the tongue surface, etc.); Obtain pulse parameters: Obtain pulse parameters through pulse diagnosis equipment or manual pulse diagnosis. Pulse parameters include the speed, strength and fluctuation shape of the pulse. Establishing a constitution identification model: Based on the symptom feature vector, tongue parameters and pulse parameters, the constitution identification model is established using the support vector machine algorithm; Constitution type identification: The user's symptom feature vector, tongue parameters and pulse parameters are analyzed through the constitution identification model to identify the user's constitution type, and the corresponding associated symptom group is output according to the constitution type. The associated symptom group includes symptoms related to the constitution type, which serves as the basis for subsequent symptom association analysis; Output constitution type and associated symptom groups: Output the user's constitution type and associated symptom groups based on the constitution identification results. Constitution types include Yin deficiency, Yang deficiency, and Qi deficiency. The constitution identification model determines the corresponding symptom groups based on different constitution types to ensure the personalization and pertinence of the symptom groups. The TCM constitution identification module accurately identifies the user's constitution type by comprehensively analyzing the user's symptom feature vector, tongue image and pulse parameters, and outputs the symptom group related to the constitution type. This module provides a scientific basis for subsequent symptom association analysis and personalized consultation path generation, ensuring the accuracy and personalization of diagnosis.

[0023] Quantification of tongue and pulse parameters Tongue quality: red (code 1), pale (code 2), tender (code 3); Pulse condition: speed (heart rate value), strength (pressure sensor reading), fluctuation shape (waveform spectrum analysis results).

[0024] Support vector machine model training: Input: multidimensional feature vector (symptoms, tongue image, pulse parameters); Output: Constitution type label (Yin deficiency, Yang deficiency, Qi deficiency); Kernel function selection: radial basis function (RBF) kernel, penalty parameter C=1.0, and error tolerance ε=0.1.

[0025] Symptom association analysis module Receiving constitution type and symptom group: Obtaining the user's constitution type (yin deficiency, yang deficiency or qi deficiency) and its corresponding associated symptom group through the TCM constitution identification module. Each constitution type has its corresponding symptom group, which includes symptoms and their attributes related to the constitution type; Calling the knowledge graph: According to the user's constitution type, the corresponding TCM symptom knowledge graph is called. The TCM symptom knowledge graph includes various TCM symptoms and their relationship with constitution type and other symptoms; Each node in the knowledge graph represents a symptom, and each edge represents the relationship between symptoms, including information such as intensity and similarity; Multi-dimensional weight calculation: Based on symptom groups and knowledge graphs, multi-dimensional weight calculation is performed. For each symptom in each symptom group, a weight is assigned to each symptom based on the following factors: Correlation between symptoms and physical constitution: According to the correlation between physical constitution and symptoms, the basic weight of each symptom is calculated, which is set by historical data, medical literature or expert knowledge; Similarity between symptoms: Calculate the similarity between symptoms and use cosine similarity to measure the similarity of symptoms in their manifestations; The joint relationship between symptoms and other symptoms: According to the connection relationship between symptoms in the knowledge graph, the joint probability or correlation between symptoms is calculated and the weight is adjusted. Strong associations between symptoms will increase their weight in the final sequence; Feedback on user input information: adjust the weight according to the intensity and change of the user's input symptoms, making the calculation of symptom correlation more personalized; Generate a high-correlation symptom sequence: Based on the calculated weights, a high-correlation symptom sequence is generated by sorting by weight. The high-correlation symptom sequence includes symptoms related to the user's physical type, which are arranged from high to low in terms of correlation. Based on the sorting results, the top 20% of symptoms with the highest correlation are selected as the high-correlation symptom sequence to ensure the personalization and targeting of the symptoms. This symptom sequence can be used for subsequent generation of personalized consultation paths.

[0026] Output symptom sequence: The generated highly correlated symptom sequence is provided to the personalized consultation path generation module for building a dynamic interactive consultation tree. This sequence ensures that the symptom check during the consultation process can cover the relevant symptoms of the user's physical type and conduct an in-depth investigation.

[0027] Knowledge graph call: Match the preset TCM symptom knowledge graph according to the constitution type, for example: Symptoms associated with Yin deficiency constitution: dry mouth, hot flashes, night sweats; Knowledge graph edge weights: association strength between symptoms (0-1), joint probability (e.g., the joint probability of "dry mouth" and "red tongue" is 0.85).

[0028] Weight calculation rules: Correlation weight: The basic value is set according to medical literature (e.g. the correlation between "dry mouth" and "yin deficiency" is 0.9); User feedback weight: If the user feedback indicates that the degree of “dry mouth” has increased, the weight will be increased by 10%.

[0029] Personalized consultation path generation module Receiving highly correlated symptom sequences: The personalized consultation path generation module receives highly correlated symptom sequences generated by the symptom correlation analysis module, which include symptoms related to the user's physical type and with a high degree of correlation. The highly correlated symptom sequences are arranged from high to low in terms of correlation, ensuring the priority and importance of the symptoms.

[0030] Determine the mandatory symptom verification items: Based on the highly correlated symptom sequence, the personalized consultation path generation module selects the symptoms with the highest symptom correlation from the sequence as mandatory symptom verification items. Mandatory symptom verification items refer to those symptoms that are crucial to determining the user's physical type, and these symptoms must be verified during the consultation process. The selected mandatory symptom verification items have the following characteristics: The most associated symptoms; High certainty, which can help doctors make a clear diagnosis; It is highly correlated with the physical type and can accurately reflect the user's health status.

[0031] Define optional symptom extensions: Based on the remaining symptoms in the highly correlated symptom sequence, the personalized consultation path generation module selects the first 20% of the remaining symptoms as optional symptom extensions. The optional symptom extensions are used to further refine and supplement the user's symptom information, helping doctors to more comprehensively assess the user's health status. The selected optional symptom extensions have the following characteristics: The correlation is lower but still somewhat related to physical constitution type; It has a certain degree of personalization and can flexibly choose whether to verify according to the user's specific situation; It can help doctors identify potential health problems that may not appear.

[0032] Constructing an interactive consultation tree: Based on the mandatory symptom verification items and optional symptom extension items, the personalized consultation path generation module generates an interactive consultation tree through a dynamic branching algorithm; The structure of the consultation tree includes multiple nodes, each node represents a symptom or symptom group, and the branches represent the relationship between symptoms and the consultation process. The specific steps are as follows: Starting from the root node, the root node represents the user's physical type and preliminary symptom analysis; Each branch node represents a symptom or a group of symptoms, and the branches connecting different nodes represent the dependencies between symptoms; Mandatory symptom verification items appear as core nodes at key nodes of the consultation tree to ensure that these symptoms are verified first during the consultation process; Optional symptom expansion items appear as branches or child nodes. After verifying the mandatory symptoms, if conditions permit, they continue to expand downward to further improve the symptom information; The connection relationship between each node is based on the correlation between symptoms and the actual needs of doctors, and the branch generation rules are dynamically adjusted.

[0033] Dynamically adjust the consultation path: The personalized consultation path generation module dynamically adjusts the structure and branches of the consultation tree based on the user's real-time feedback and symptom occurrence. The specific adjustment basis includes: Symptom changes: If a symptom shows new changes during the consultation process, the consultation tree will automatically adjust to add new verification items or expansion items; User feedback: Dynamically change the order of the consultation path or add new symptom items based on user feedback on symptoms (such as symptom intensity, frequency, etc.); Doctor input: Based on the doctor's judgment, further refine the consultation path and adjust the priority between required and optional items.

[0034] Output personalized consultation path: Based on the generated interactive consultation tree, the personalized consultation path generation module outputs a personalized consultation path, which provides doctors with a systematic and personalized consultation process, ensuring that the consultation process is both efficient and targeted, helping doctors quickly identify the user's core symptoms and make an accurate preliminary diagnosis.

[0035] Based on the mandatory symptom verification items and optional symptom extension items, the personalized consultation path generation module generates an interactive consultation tree through a dynamic branching algorithm, including: Define the interactive consultation tree structure: The interactive consultation tree structure is defined as a hierarchical tree structure, including multiple nodes and branches. The root node of the tree represents the starting point of the consultation (i.e., the user's physical type and preliminary symptom analysis), the nodes represent symptoms, symptom groups or further verification items of symptoms, and the branches represent the relationship, priority and dependency conditions between symptoms; The basic structure includes: Root node: includes the user's physical type and preliminary symptom information; Mandatory symptom verification items: Symptoms with high correlation (such as symptoms A, B, and C) are used as core nodes in the consultation path; Optional symptom expansion items: Use symptoms with lower correlation but still certain relevance (such as symptoms D, E, and F) as branch nodes, relying on mandatory symptom verification items or user feedback.

[0036] Generate a preliminary consultation tree based on mandatory symptoms: The personalized consultation path generation module generates the core part of the consultation tree based on the mandatory symptom verification items. These symptoms are usually key symptoms that must be verified at the beginning of the user's consultation, so they will be set as key nodes in the consultation tree and placed under the child nodes of the root node. The structure of the consultation tree is initially formed: Root node: constitution type information (such as Yin deficiency, Yang deficiency, Qi deficiency); Mandatory symptom verification item node: Each symptom is added as a child node to the root node; For example: Node 1: Symptom A (e.g., “dry mouth”), Node 2: Symptom B (e.g., “tiredness”), etc.

[0037] Selectively add optional symptom extensions: Based on the remaining symptoms in the highly correlated symptom sequence, the personalized consultation path generation module selects optional symptom extensions as sub-branches through a dynamic branching algorithm and connects to relevant nodes of the consultation tree. These symptoms are selected based on the feedback of the mandatory symptom verification items or the correlation of the symptom groups.

[0038] Dynamic judgment: Based on the user's feedback or symptom manifestations, the dynamic branching algorithm determines whether to add optional symptom extensions to the consultation path. If the user's symptoms meet the extension conditions, the relevant symptoms are added to the tree as new branches or child nodes. For example: If the user reports that symptom A "dry mouth" is severe, the system may guide the user to further consultation on symptom C (such as "red tongue").

[0039] If symptom A, “dry mouth,” is mild, the system may skip symptom C or, depending on the user’s response, direct the inquiry to symptom D (e.g., “constipation”).

[0040] Dynamically adjust branch generation rules: During the interactive consultation process, the consultation tree is generated based on the required symptoms and optional symptom extensions, and is dynamically adjusted based on the user's real-time feedback. The dynamic branching algorithm ensures that the consultation tree can adjust its structure in real time according to changes in user symptoms. The specific steps are as follows: User feedback: Dynamically adjust the path based on the user's answer, the intensity and frequency of the main symptoms, etc. For example, if the user reports that a symptom has changed during the consultation (such as symptom A "dry mouth" has worsened), the system will automatically adjust and add extensions related to the symptom (such as symptom C or symptom F); Symptom dependency: Based on the knowledge graph or empirical data, symptoms may have certain associations or dependencies. The system updates the branches of the consultation tree in real time based on these dependencies. For example, if symptom A is highly correlated with symptom B, when symptom A is confirmed, the system will automatically expand further consultation on symptom B; User symptom evolution: If the user shows an evolution of multiple symptoms during the initial consultation, the dynamic branching algorithm can help adjust the consultation tree to add other optional symptom items to form a more precise personalized path.

[0041] Build dynamically adjustable consultation nodes: Through the dynamic branching algorithm, each node dynamically adjusts the content and structure according to the needs of the real-time consultation process: Expandable nodes: For example, a "Symptom A" node may include multiple child nodes and dynamically generated expansion items (e.g., "Symptom B", "Symptom C"); Jump logic: Based on the judgment of the algorithm, the system can skip certain nodes and jump directly to the next symptom verification item to ensure the efficiency and pertinence of the consultation; Output the final interactive consultation tree: The personalized consultation path generation module outputs the generated interactive consultation tree to form a multi-level, dynamically adjusted consultation process, which includes: Mandatory symptom verification items: ensure that the most critical symptoms are verified first during the consultation process; Optional symptom expansion: Based on user feedback, further expand symptom-related information to improve the comprehensiveness of the consultation; Dynamically adjust the path: Ensure that the consultation path can be flexibly adjusted as user feedback changes.

[0042] The personalized consultation path generation module generates an interactive consultation tree based on the required symptom verification items and optional symptom expansion items through a dynamic branching algorithm. During the entire process, the system will dynamically adjust the consultation path based on the user's real-time feedback and symptom changes, making the consultation process both efficient and flexible to respond to various clinical situations, thereby providing doctors with an accurate symptom verification path.

[0043] Dynamic branching algorithm logic: Mandatory symptom verification item: mandatory verification of symptoms (such as "dry mouth"), if not passed, the path is terminated; Optional symptom expansion: selectively expand according to the user's answer (e.g. after confirming "dry mouth", jump to the "red tongue" or "constipation" branch); Real-time adjustment: If the user reports that "dry mouth" has been relieved, the system will automatically skip the sub-symptoms with reduced relevance.

[0044] Data integration and medical record generation module Obtaining interactive data of the consultation tree: The data integration and medical record generation module obtains interactive data of the interactive consultation tree from the personalized consultation path generation module. The interactive data includes: The user's answers during the consultation process (such as the manifestation, severity, duration, etc. of symptoms); Selection of mandatory symptom verification items and optional symptom expansion items; Selection and jump records of paths in the question tree.

[0045] Obtaining historical medical records: The data integration and medical record generation module extracts the user's historical medical records from the hospital database, which includes: The user's medical history (such as past physical type, common symptoms, diagnosis and treatment plans, etc.); past diagnoses, treatments and their effectiveness; Previous constitution type and symptom evolution history; The purpose of obtaining historical medical records is to provide complete background information for the generation of current medical records and to ensure the consistency and comprehensiveness of medical records; Integration of interactive data and historical medical records: After obtaining the interactive data and historical medical records of the consultation tree, the data integration and medical record generation module integrates the interactive data and historical medical records. The integration process includes: Match symptoms and constitution types: Compare the symptoms obtained in the current consultation tree with previous symptoms in historical diagnosis and treatment records, and update the symptom evolution path.

[0046] Constitution type update: Based on the current user's symptoms and feedback, combined with the constitution type in the historical records, the user's constitution feature mark is updated. For example, if the current consultation results match the historical symptoms, the user's constitution type mark may be updated to Yin deficiency, Yang deficiency or Qi deficiency.

[0047] Symptom evolution record: Combine historical symptom records with current symptom data to generate a symptom evolution path, and integrate it into the pre-diagnosis medical record to record the changing trend of symptoms.

[0048] Generate pre-diagnosis medical records including physical characteristics markers: Based on the integrated interactive data and historical diagnosis and treatment records, the data integration and medical record generation module generates pre-diagnosis medical records, which include the following contents: Physical characteristics marking: Mark the user's physical type (such as Yin deficiency, Yang deficiency or Qi deficiency) based on the current consultation results and historical diagnosis and treatment records; Symptom evolution path: describes the path of symptom change from the initial stage to the current state, reflecting the time, manifestation and related factors of symptom change; Diagnosis results: including the verified diagnosis results of mandatory symptom verification items and optional symptom extension items, as well as symptom groups related to physical constitution types; Historical information: Summarize relevant past symptoms, treatment methods and their effects in historical medical records as a reference for the current diagnostic process.

[0049] Output complete pre-diagnosis medical records: The data integration and medical record generation module will output the generated pre-diagnosis medical records to doctors or relevant medical personnel. The output pre-diagnosis medical records include complete physical characteristic markers, symptom evolution path and historical diagnosis and treatment information, providing a scientific basis for further diagnosis and treatment process.

[0050] The data integration and medical record generation module generates a complete pre-diagnosis medical record including physical characteristic markers and symptom evolution path by integrating the interactive data from the personalized consultation path generation module and the user's historical diagnosis and treatment records. This process ensures the consistency and comprehensiveness of the medical record, which helps to provide doctors with more accurate diagnosis and treatment plans. At the same time, a feedback optimization mechanism is introduced in the medical record generation process, so that the medical record content can be continuously optimized and updated as the diagnosis and treatment process progresses.

[0051] Feedback Optimization Module Obtaining the difference characteristics between clinical diagnosis data and pre-diagnosis medical records: The feedback optimization module obtains the difference characteristics between clinical diagnosis data and pre-diagnosis medical records. The steps are as follows: Obtain clinical diagnosis data: Obtain the final diagnosis results confirmed by the doctor from the clinical diagnosis system, including: Clinically diagnosed symptoms, constitution types and their associated symptoms; Treatment plan and effect after diagnosis.

[0052] Compare pre-diagnosis medical records: Compare the symptoms, constitution types, and symptom evolution paths generated in the pre-diagnosis medical records based on user feedback and the interactive question tree with the actual clinical diagnosis data to find out the differences between the two, for example: Certain symptoms in the preliminary medical history were not diagnosed or the diagnosis results were different; There is a difference between the constitution type marker and the constitution type finally diagnosed.

[0053] Identify the correlation between differential features and knowledge graph: After identifying the differential features in the pre-diagnosis medical records and clinical diagnosis data, associate the differential features with the content in the knowledge graph to determine the parts that need to be adjusted. The steps include: Extraction of differential features: Differential features include information such as the severity of symptoms, correlation, and judgment bias of physical type. The module extracts key differential features from confirmed data and medical records; Knowledge graph matching: Match these difference features with the existing symptoms, constitution types, symptom correlation, and the relationship between constitution and symptoms in the knowledge graph to determine which knowledge graph entries have deviations in the current medical records or need to be optimized; For example, if the correlation between the symptom of "dry mouth" and "yin deficiency" in the pre-diagnosis medical record is low, but the clinical diagnosis shows that the correlation between the symptom and "yin deficiency" is high, then the correlation between the symptom of "dry mouth" and "yin deficiency" needs to be adjusted; Adjust the weight coefficient of the knowledge graph: After identifying the difference features and matching them with the relevant content of the knowledge graph, adjust the weight coefficient of the knowledge graph through the parameter self-correction unit. The steps are as follows: Weight coefficient adjustment: Automatically adjust the weight coefficient between symptoms and constitution types in the knowledge graph based on the difference characteristics and matching results. For example: If certain symptoms are more prominent in clinical diagnosis but not given enough attention in pre-diagnosis medical records, the feedback optimization module will increase the weight between the symptom and the relevant constitution type; If some symptoms are too highly correlated with the physical type but have a low correlation in the actual diagnosis, the feedback optimization module will reduce the weight coefficients of these symptoms; Knowledge graph update: After each adjustment of the weight coefficient, the knowledge graph will be updated so that a more accurate correlation between symptoms and constitution can be referenced in the subsequent consultation process. This adjustment enables future pre-diagnosis medical records to better match clinical diagnosis results; Feedback optimization effect evaluation: By evaluating the effect of the adjusted knowledge graph, the feedback optimization module continuously monitors the performance of the optimized knowledge graph in the process of consultation and constitution identification; Monitor optimization results: In the new round of user consultation process, monitor whether the optimized knowledge graph can effectively reduce the difference with the clinical diagnosis results.

[0054] Feedback correction: If the optimization results are not ideal, the feedback optimization module will further analyze the errors or deviations and conduct a second round of adjustments to ensure that the final optimized knowledge graph can more accurately support constitution identification and symptom association analysis.

[0055] Generate and output the optimized knowledge graph: The optimized knowledge graph will be used in subsequent medical consultation and constitution identification and output to the system to ensure the accuracy and rationality of the optimized knowledge graph in the process of generating interactive medical consultation trees and symptom association analysis; The feedback optimization module achieves dynamic optimization of the knowledge graph by obtaining the difference characteristics between clinical diagnosis data and pre-diagnosis medical records, and adjusting the weight coefficient of the knowledge graph through the parameter self-correction unit. This process can not only reduce the deviation in system diagnosis, but also improve the accuracy of the system's recognition of user symptoms and physical types, thereby providing more accurate data support for the subsequent consultation process.

[0056] Parameter self-correction unit: Extraction of differential features: If the clinical diagnosis is "Yang deficiency" but the pre-diagnosis is "Qi deficiency", the deviation of constitution type is marked; Weight adjustment: Increase the association weight between the symptom of "fear of cold" and "yang deficiency" from 0.7 to 0.9; Effect evaluation: After optimization, the system's accuracy in identifying the physical constitution of new user data increased by 15%.

[0057] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0058] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0059] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A TCM pre-consultation collection and management system, characterized in that: The system comprises: The user information collection module is used to collect basic user information and main symptoms, and convert them into symptom feature vectors to generate an initial data set; The TCM constitution identification module is used to construct a constitution identification model based on symptom feature vectors, tongue parameters and pulse parameters through the support vector machine algorithm, and output the user's constitution type and associated symptom groups; The symptom association analysis module is used to call the TCM symptom knowledge graph according to the user's constitution type and generate a highly correlated symptom sequence through multi-dimensional weight calculation; A personalized consultation path generation module is used to generate an interactive consultation tree containing mandatory symptom verification items and optional symptom extension items based on a highly correlated symptom sequence through a dynamic branching algorithm, and dynamically adjust the path based on user feedback; The data integration and medical record generation module is used to integrate the interactive consultation tree interaction data and historical diagnosis and treatment records to generate a pre-diagnosis medical record containing physical characteristic markers and symptom evolution paths; The feedback optimization module is used to compare the difference characteristics between clinical diagnosis data and pre-diagnosis medical records, and adjust the knowledge graph weight coefficient to optimize the system's diagnostic accuracy.

2. The system according to claim 1, characterized in that The specific workflow of the user information collection module is as follows: Obtain user basic information and main symptoms through interactive interface and form input; Convert the user's basic information and main symptoms into a numerical symptom feature vector containing physical information, main symptoms and background information; The symptom feature vector and basic information are integrated to generate the initial data set, which is then stored in the system database.

3. The system according to claim 1, characterized in that In the TCM constitution identification module: Tongue parameters include quantitative data of tongue quality, tongue coating, and tongue shape; Pulse parameters include numerical parameters of pulse speed, strength and fluctuation shape; The support vector machine algorithm adopts a radial basis function kernel and trains a constitution identification model through a multi-dimensional feature vector to identify the constitution types of Yin deficiency, Yang deficiency and Qi deficiency.

4. The system according to claim 3, characterized in that The construction of the constitution identification model specifically includes: Integrate the user's symptom feature vector, tongue parameters and pulse parameters into a comprehensive multi-dimensional feature vector; The user's physical constitution type is divided into Yin deficiency, Yang deficiency and Qi deficiency as classification labels; Construct a constitution identification model through support vector machine algorithm; The multi-dimensional feature vector and the corresponding classification label are used as input to train the constitution identification model; Use the trained constitution identification model to predict the constitution type of new user data.

5. The system according to claim 1, characterized in that In the symptom association analysis module: The multidimensional weight calculation includes the correlation between symptoms and constitution types, the similarity between symptoms, the joint relationship and the user feedback weight; The highly correlated symptom sequences are sorted by weight, and the first 20% of the symptoms are defined as mandatory symptom verification items, and the first 20% of the remaining symptoms excluding the mandatory symptom verification items are defined as optional symptom extension items.

6. The system according to claim 1, characterized in that In the personalized consultation path generation module: The dynamic branching algorithm generates the core nodes of the question tree based on the mandatory symptom verification items, and selectively adds optional symptom extension items as sub-branches; The interactive consultation tree nodes dynamically adjust the content and structure according to real-time user feedback to achieve path jumping and expansion.

7. The system according to claim 6, characterized in that The interactive diagnosis tree including mandatory symptom verification items and optional symptom extension items is generated by the dynamic branching algorithm, specifically including: The interactive consultation tree structure is defined as a hierarchical tree structure, including several nodes and branches. The root node of the tree represents the starting point of the consultation, the nodes represent the symptoms, and the branches represent the relationship, priority and dependency conditions between the symptoms. Generate the core part of the medical inquiry tree according to the mandatory symptom verification items, the core part including the root node and the mandatory symptom node; According to the remaining symptoms in the highly correlated symptom sequence, the optional symptom extension items are selected as sub-branches through the dynamic branching algorithm and connected to the consultation tree; During the interactive consultation process, an interactive consultation tree is generated based on the mandatory symptom verification items and optional symptom extension items, and is dynamically adjusted based on real-time feedback from users; Through the dynamic branching algorithm, the content and structure of each node are dynamically adjusted according to the needs of the real-time consultation process; The generated interactive consultation tree is output to form a multi-level and dynamically adjusted consultation process.

8. The system according to claim 7, characterized in that The specific workflow of the data integration and medical record generation module is as follows: Match the interactive consultation tree interaction data with the previous symptoms, physical constitution types and treatment results in the historical diagnosis and treatment records; Generate structured pre-diagnosis medical records containing constitution type markers, symptom evolution paths and historical information.

9. The system according to claim 1, characterized in that In the feedback optimization module: The differential features include symptom association bias and misjudgment of constitution type, and the knowledge graph weight is adjusted through the parameter self-correction unit; The optimized knowledge graph is used to update the operational logic of the symptom association analysis module and the constitution identification module.

10. The system according to claim 1, characterized in that The TCM symptom knowledge graph is a dynamic graph, in which nodes represent symptoms, edges represent correlations and weight coefficients between symptoms, and adaptive weight adjustment is achieved through a feedback optimization module.

Images