User Health Intervention Feature Management System and Method Based on Multi-Source Data

Abstract

This invention discloses a user health intervention feature management system and method based on multi-source data, belonging to the field of medical information technology. The method includes: verifying user identity and connecting to a medical management platform to obtain medical data (including medical time, diagnosis, and medication); constructing a supplementary extended database containing non-medical and non-platform medical databases; integrating relevant data to form a user-coordinated dataset; parsing the dataset to obtain a user medical feature library; opening a verification feedback interface based on the feature library; and managing and adjusting the user medical database based on feedback data. This invention solves the technical problems of fragmented, incomplete, and difficult-to-achieve personalized and precise services from existing user medical data, achieving the technical effect of effectively managing user medical data and providing continuous high-quality data support for medical services.

Description

Technical Field

[0001] This application relates to the field of medical information technology, and in particular to a user health intervention feature management system and method based on multi-source data. Background Technology

[0002] In the field of pharmaceutical data management, the comprehensive and accurate acquisition and effective utilization of user medical data are becoming increasingly prominent issues. Efficiently integrating and deeply analyzing pharmaceutical data to meet the needs of personalized medical services has become key to driving the development of pharmaceutical data management. Traditional pharmaceutical data management methods typically have many limitations and shortcomings. They rely solely on specific medical management platforms to collect data, neglecting user self-purchase records and non-platform medical data, resulting in fragmented and incomplete data. Furthermore, they lack depth and accuracy in analyzing user health and intervention characteristics, making it difficult to fully extract data value and provide strong support for medical decision-making. In terms of data fusion and utilization, the methods are simplistic and inefficient, failing to adapt to the increasingly complex needs and changes in the medical field.

[0003] At present, the relevant technologies face technical challenges such as fragmented and incomplete user medical data, making it difficult to achieve personalized and precise services. Summary of the Invention

[0004] This application provides a user health intervention feature management system and method based on multi-source data. It verifies user identity and connects to a medical management platform to obtain medical data (including medical time, diagnosis, and medication). It constructs a supplementary extended database including non-medical and non-platform medical databases, integrates relevant data to form a user-coordinated dataset, parses this dataset to obtain a user medical feature library, opens a verification feedback interface based on the feature library, and manages and adjusts the user medical database based on feedback data. By comprehensively integrating multi-source heterogeneous medical data, it deeply mines and accurately analyzes user medical features, achieving effective management of user medical data and providing continuous high-quality data support for medical services.

[0005] This application provides a user health intervention feature management system based on multi-source data, including:

[0006] The system comprises the following modules: a medical data acquisition module, which verifies user identity information through an authorized channel, connects to a health intervention management platform based on the verified user identity information, and acquires a user health intervention time-series dataset, including platform intervention records, non-platform intervention records, and self-managed health behavior records; a supplementary and extended database construction module, which parses unstructured health records provided by the user, generates standardized data, constructs a supplementary and extended database, and aligns non-health intervention management platform intervention data within it using a knowledge graph; a unified dataset construction module, which integrates the time-series dataset, the unstructured health records, and non-health intervention management platform intervention data to construct a user unified dataset; a medical feature library acquisition module, which analyzes user health features and intervention features based on the user unified dataset to obtain a user health intervention feature map; and a management and adjustment module, which opens a verification feedback interface based on the user health intervention feature map. This verification feedback interface connects to the user terminal to annotate and provide feedback on features in the user health intervention feature map, and manages and adjusts the user health intervention feature map based on the feedback data obtained from the verification feedback interface.

[0007] This application provides a method for managing user health intervention features based on multi-source data, including:

[0008] User identity information is verified through an authorized channel. Based on the verified user identity information, a health intervention management platform is connected to obtain a user health intervention time-series dataset, which includes platform intervention records, non-platform intervention records, and self-managed health behavior records. Unstructured health records provided by the user are parsed to generate standardized data, and a supplementary extended database is constructed. Non-health intervention management platform intervention data is aligned using a knowledge graph. The time-series dataset, unstructured health records, and non-health intervention management platform intervention data are fused to form a user-coordinated dataset. User health features and intervention features are analyzed based on the user-coordinated dataset to obtain a user health intervention feature map. A verification feedback interface is opened based on the user health intervention feature map. This interface connects to the user terminal to annotate and provide feedback on the features in the user health intervention feature map. The user health intervention feature map is managed and adjusted based on the feedback data obtained from the verification feedback interface.

[0009] The proposed user health intervention feature management system and method based on multi-source data first verifies user identity and connects to a medical management platform to obtain medical data (including medical time, diagnosis and treatment, and medication). It then constructs a supplementary extended database that includes non-medical and non-platform medical databases, integrates relevant data to form a user-coordinated dataset, parses the dataset to obtain a user medical feature library, opens a verification feedback interface based on the feature library, and manages and adjusts the user medical database based on the feedback data. This achieves the technical effect of effectively managing user medical data and providing continuous high-quality data support for medical services. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0011] Figure 1 A schematic diagram of the structure of a user health intervention feature management system based on multi-source data provided in this application embodiment;

[0012] Figure 2 This is a flowchart illustrating the user health intervention feature management method based on multi-source data provided in this application embodiment.

[0013] Figure labeling: Medical data acquisition module 10, supplementary and extended database construction module 20, unified dataset construction module 30, medical feature database acquisition module 40, management and adjustment module 50. Detailed Implementation

[0014] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0016] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. 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 belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0017] This application provides a user health intervention feature management system based on multi-source data, such as... Figure 1 As shown, the method includes:

[0018] The medical data acquisition module 10 is used to verify user identity information through an authorized channel. Based on the verified user identity information, it connects to the health intervention management platform to obtain a time-series dataset of user health interventions. This time-series dataset includes platform intervention records, non-platform intervention records, and self-managed health behavior records. Specifically, the medical data acquisition module 10 performs a comprehensive identity verification process on user operation requests, identifying key identity information such as the user's input ID card number, set password, and scanned QR code. It uses data comparison and verification technology to confirm the legality of the user's identity and analyzes the logical relationships and consistency between user identity information. Based on the identity verification results, it determines the validity of the user's identity. The validity determination information includes whether the identity verification was successful and possible identity anomalies (such as information mismatch, exceeding the verification limit, etc.), which is used to decide whether to establish a connection with the medical management platform. The obtained identity validity determination information is used to extract and analyze data features. These data features refer to key factors affecting the acquisition of medical data, such as identity verification type and verification rules. Through analysis algorithms, identity determination cases with similar data features are grouped together. Based on the analysis results and the characteristics of the user's identity, a decision is made on whether to obtain the corresponding user medical data. Each category of user medical data represents the medical information of a group of users with the same or similar identity characteristics.

[0019] In one possible implementation, the medical data acquisition module 10 includes an identity verification rule setting unit. This unit sets identity verification rules, including identity verification type, verification quantity, and verification order. The identity verification type includes ID card, password, and QR code. Specifically, the identity verification rule setting unit allows users to flexibly set identity verification rules according to actual circumstances. Users can choose one of the following as a single verification method: ID card, password, or QR code. For example, they can choose only password verification. To improve security, users can also choose a combination of multiple verification methods, such as scanning the ID card first and then entering the password. Users can also determine the number of verifications, such as setting two verifications, and specify the verification order, such as QR code verification first, then password verification.

[0020] The verification module construction unit is used to extract verification features according to the authentication rules, establish the temporal relationship between the authentication type and the authentication order, and fit and construct a user verification module. Specifically, the verification module construction unit extracts key verification features according to the rules determined by the authentication rule setting unit. For example, if the rule sets that ID card scanning is performed before password input, the unit extracts the features of ID card scanning and password input respectively, establishes the temporal relationship between the authentication type (such as ID card scanning, password input) and the authentication order (scanning before input), and formulates a precise sequence of operation steps. Based on this, a verification module specifically adapted to the user settings is constructed through fitting.

[0021] The verification result acquisition unit is used to sequentially verify the user verification module according to the time sequence to obtain the verification result. Specifically, after the verification module is constructed, the verification result acquisition unit will perform verification operations sequentially according to the time sequence specified in the user verification module. If the verification order is ID card scanning first and then password input, then the ID card scanning result will be judged first. If it passes, then the password input verification will be performed to obtain the verification result, that is, to determine whether the user's identity verification is successful or unsuccessful.

[0022] The communication connection unit is used to generate authorization information when the verification result is successful, thereby establishing a communication connection with the health intervention management platform. Specifically, when the verification result acquisition unit obtains a successful verification result, the communication connection unit starts working, generates corresponding authorization information, and establishes a secure and reliable communication connection with the medical management platform based on this authorization information, thus preparing for subsequent operations such as obtaining user medical data.

[0023] The supplementary extended database construction module 20 is used to parse unstructured health records provided by users, generate standardized data, construct a supplementary extended database, and align non-health intervention management platform intervention data within it through a knowledge graph. Specifically, the supplementary extended database construction module 20 clarifies the task objectives of constructing the supplementary extended database, creating a comprehensive database structure that includes both non-medical treatment databases and non-platform medical treatment databases. For the non-medical treatment database, the supplementary extended database construction module 20 collects and organizes user-purchased medication records through various means, such as connecting with pharmacy sales systems to obtain user purchase lists; or providing an interface for users to upload their own medication purchase records, allowing users to actively submit their medication information. The acquired data is standardized to ensure consistent format and accuracy. For the non-platform medical treatment database, the supplementary extended database construction module 20 collects other data obtained from the current medical management platform... Medical data involves data interaction with other medical institutions' systems to obtain information such as patients' medical records, diagnostic reports, and treatment plans from other hospitals. It may also involve cooperation with third-party medical data providers to obtain relevant data resources. After acquiring this data, it undergoes careful screening, organization, and transformation to conform to the format and requirements of the supplementary extended database. After completing the above data collection and processing of non-medical databases and non-platform medical databases, the supplementary extended database construction module 20 integrates the data to build a complete supplementary extended database. This database complements the user medical data obtained from the medical management platform, providing more comprehensive and richer information support for subsequent data analysis, feature extraction, and medical services.

[0024] In one possible implementation, the supplementary extended database construction module 20 includes: an interface creation unit, which is used to create a data addition interface for scanning or entering user unstructured records. Specifically, the interface creation unit creates a data addition interface designed to meet the needs of scanning or directly entering user medical and pharmaceutical purchase records. For example, it can support scanners to quickly scan users' paper purchase receipts or medical records, converting them into electronic data. Simultaneously, it provides an intuitive input interface, allowing operators to manually enter relevant information. Whether it's a user's detailed treatment process at the hospital or their purchase details at the pharmacy, this convenient interface allows for accurate and error-free entry into the system.

[0025] An added data identification unit is used to identify added data using identification technology, determining multi-dimensional identification information of the added data, including intervention plan, implementation time, and physiological indicators. Specifically, the added data identification unit analyzes and identifies newly added data, accurately determining multi-dimensional identification information from complex data, such as accurately capturing the time of medical treatment, and extracting at least one key piece of information from diagnostic and medication data. This identification provides an accurate and valuable data foundation for subsequent processing.

[0026] The data deduplication unit is used to perform similarity calculations between the multi-dimensional identification information and the user health intervention time-series dataset to determine newly added data. Specifically, the data deduplication unit uses the previously acquired multi-dimensional identification information to compare and check existing user medical data one by one. Through comparison, it can accurately identify which data is newly added, ensuring the novelty and accuracy of the data and avoiding duplication and redundancy.

[0027] A type identification unit is used to identify the type of the added data and align non-medical management platform intervention data through a knowledge graph. Specifically, the type identification unit analyzes the newly added data to accurately distinguish its type. Data related to users purchasing medication on their own is identified as non-medical data and will be properly stored in the non-medical database. Medical data from outside the platform, such as treatment records from clinics or traditional Chinese medicine practitioners, is identified as non-platform medical data and stored accordingly in the non-platform medical database. This categorized storage ensures that all types of data are properly managed and utilized.

[0028] The integrated dataset construction module 30 is used to integrate the time-series dataset, the unstructured health records, and intervention data from non-health intervention management platforms to construct a user integrated dataset. Specifically, the integrated dataset construction module 30's task is to integrate user-related data from different sources to obtain user medical data. This data includes detailed medical information from formal medical platforms, such as the time of treatment, the specific treatment process, and the medications used. The integrated dataset construction module 30 also accesses non-medical databases, which include records of users purchasing medications themselves, reflecting their self-medication practices. Furthermore, the integrated dataset construction module 30 incorporates non-platform medical databases, collecting other medical data obtained by users outside of medical management platforms, such as treatment experiences at small clinics and traditional Chinese medicine clinics. After classifying the data, the unified dataset construction module 30 performs data fusion, standardizing data from different sources, in different formats and structures to ensure a unified format and standard. Through data matching and association technologies, it links users' medical treatment and medication purchase data in different scenarios to form a comprehensive and coherent whole. For example, it links users' hospital treatment data with their self-purchased medication records to analyze whether there are any supplementary interventions or conflicts. Then, it combines non-platform medical treatment data with platform medical treatment data to fully present the user's medical history, successfully constructing a unified user dataset, providing comprehensive, accurate and integrated basic data for subsequent analysis and applications.

[0029] The medical feature database acquisition module 40 is used to analyze user health characteristics and intervention characteristics based on the user's overall data set to obtain a user health intervention feature map. Specifically, the medical feature database acquisition module 40 receives the user's overall data set, which integrates a collection of various medical-related data of the user. It mines information from the user's overall data set. In analyzing user health characteristics, it carefully analyzes the user's medical visit time patterns, disease symptoms and disease progression trends reflected in the treatment data, and other factors that may affect health, such as lifestyle habits and genetic history. For the analysis of intervention characteristics, it focuses on the type of drug, frequency of use, dosage, intervention cycle, and the combination and interaction between different drugs. It uses association rule mining algorithms to find potential associations between drugs, such as discovering that certain drugs are often used simultaneously to enhance treatment effects. Using time series analysis models, it analyzes the change patterns of intervention over time, such as studying the adjustment patterns of drug dosage at different stages of disease development. Through comprehensive calculation and analysis of algorithms, it extracts key feature information. For example, it utilizes the powerful learning and predictive capabilities of neural network models to discover the user's sensitivity to certain drugs and predict potentially effective drug combinations for specific conditions. By using a logistic regression model to evaluate the impact of various intervention characteristics on treatment efficacy, the importance and weight of each drug were determined, and a user medical characteristic database that accurately reflects the user's health status and intervention characteristics was successfully obtained, providing valuable reference for subsequent medical services and management.

[0030] In one possible implementation, the medical feature database acquisition module 40 includes a standardization processing unit, which cleans and standardizes the user-managed dataset. Specifically, the standardization processing unit processes the user-managed dataset, carefully examining each record in the data examples. If there are missing medical visit times, incomplete diagnosis results, or abnormal drug data, the missing and outlier values ​​are identified and removed. Data such as drug dosages are standardized, for example, by converting drug dosages expressed in different units in the data to mg, ensuring the consistency and comparability of data in subsequent processing.

[0031] The dataset extraction unit extracts user health data and intervention record data from the preprocessed user-coordinated dataset. It constructs temporal relationships and extracts features from the user health data and intervention record data, including frequency, interval, and changes in health status or intervention. Specifically, the dataset extraction unit accurately extracts key information from the standardized dataset. For health data, it focuses on diagnostic results, statistically analyzing the frequency and time intervals of various diseases. Taking hypertension as an example, it calculates the number of diagnoses and the time span between adjacent diagnoses. For intervention records, it analyzes in detail the names, frequency of use, and dosage changes of prescribed and self-purchased medications. Through these extraction operations, it constructs a dataset that clearly reflects the user's health and intervention status.

[0032] The feature vector construction unit quantizes and encodes the extracted features to construct feature vectors. Specifically, the feature vector construction unit transforms the extracted data into computable and comparable feature vectors. For categorical features, such as different disease diagnoses, a One-Hot encoding method is used to convert them into independent vector representations. For numerical features, such as drug dosage, normalization processing is performed to unify their numerical range, facilitating subsequent calculations and comparisons.

[0033] A feature clustering unit is used to perform feature clustering based on the feature vectors to identify user health feature categories and intervention feature categories. Specifically, the feature clustering unit uses a suitable clustering algorithm, such as the K-Means algorithm, to group users according to the constructed feature vectors. During the clustering process, the cluster centers and boundaries are continuously adjusted so that users with similar features are grouped into the same category. By evaluating the clustering effect using metrics such as the silhouette coefficient, the optimal clustering result is determined, thereby identifying user groups with similar health and intervention features.

[0034] The correlation analysis unit utilizes association mining algorithms to analyze the correlation between user health characteristics and intervention characteristics. Specifically, the correlation analysis unit employs association mining algorithms and correlation coefficient analysis methods to deeply explore the intrinsic connections between user health characteristics and medication characteristics. Through association rule mining, it discovers potential associations between diseases and commonly used drugs. Using methods such as Pearson correlation coefficient, it quantitatively analyzes the strength of correlations between numerical characteristics, providing deeper insights into understanding user health and medication patterns.

[0035] The medical feature database construction unit is used to establish mapping associations between the feature vectors and the user's health feature categories and medication feature categories, and to fit the correlations to construct the user's health intervention feature map. Specifically, the medical feature database construction unit integrates and associates the rich information processed by the preceding units to create a detailed record for each user, covering their health features, medication features, cluster categories, and related association rules, etc., adding efficient indexes to the data so that it can be quickly and accurately retrieved and used when needed, providing strong support for medical services and research.

[0036] In one possible implementation, the medical feature library construction unit includes a high-order feature output subunit. This subunit constructs the original extracted feature matrix and learns the nonlinear interactions between features using a multilayer perceptron (MLP) to output new high-order features. Specifically, the high-order feature output subunit constructs the original extracted feature matrix, which contains features initially extracted from user health data and medication record data. A MLP is then used to learn the nonlinear interactions between these features. The MLP can capture complex relationships between features and, through the operations of multiple neurons and hidden layers, discovers potential patterns and associations. Finally, it outputs new high-order features, which are a deeper and more abstract expression of the original features, revealing hidden information within the data.

[0037] The feature concatenation subunit is used to concatenate the output higher-order features with the original features to obtain an augmented feature matrix. Specifically, the feature concatenation subunit concatenates the new higher-order features generated by the higher-order feature output subunit with the original extracted features, integrating the basic information contained in the original features with the deep relationships revealed by the higher-order features to form a richer and more comprehensive augmented feature matrix. The augmented feature matrix integrates feature information from different levels and dimensions, providing a more sufficient data foundation for subsequent analysis and processing.

[0038] The regularization loss function definition subunit is used to define the regularization loss function, train the regression model, and obtain the optimal weights. Specifically, the regularization loss function definition subunit defines the regularization loss function, which consists of two parts: one part is the mean squared error between the predicted and actual values, used to measure the model's prediction accuracy; the other part is the regularization term, used to control the model's complexity and prevent overfitting. For a given number of samples n and number of features p, the regularization term is calculated by summing the squared differences between the predicted and actual values ​​for each sample and adding the regularization parameter. The loss function is obtained by summing the absolute values ​​of the feature weights. The regression model is trained using a loss function, and the optimal weights are obtained by continuously adjusting the model's parameters to minimize the loss function.

[0039] A feature selection subunit is used to select features from the added feature matrix based on the regression model, filtering out features with non-zero weights. Specifically, the feature selection subunit selects features from the added feature matrix based on the regression model trained with a regularized loss function, filtering out features with non-zero weights. Features play an important role in the prediction and interpretation of the model, and can effectively reduce the dimensionality of features, improving the efficiency and generalization ability of the model. For example, when training a Lasso regression model using L1 regularization, features whose weights are compressed to zero are considered to have little impact on the results and are therefore discarded, retaining only the key features with non-zero weights, providing a more concise and effective feature set for subsequent applications.

[0040] In one possible implementation, the medical feature database acquisition module 40 includes a temporal relationship establishment unit. This unit extracts user health features and intervention features from the user medical feature database and establishes a temporal relationship between the medication features and the user health features. Specifically, the temporal relationship establishment unit acquires the user medical feature database, extracts the user's health features and intervention features from it, and analyzes the changes and interrelationships of these features over time. For example, it observes the order and time intervals of different drug use during the development of a disease, as well as the correlation between drug use and improvement or deterioration of health status, thereby establishing a clear temporal relationship between medication features and user health features.

[0041] The list determination unit is used to perform temporal characteristic analysis based on the medication characteristics to determine an intervention whitelist, a restricted intervention list, and an intervention contraindication list. The medication whitelist consists of interventions that are effective for the user's symptoms based on the temporal characteristics of the intervention. The restricted intervention list consists of drugs whose cumulative intervention time exceeds a threshold and which are ineffective or significantly less effective for the user's symptoms based on the temporal characteristics of the intervention. The intervention contraindication list consists of interventions that are ineffective for the user's symptoms based on the temporal characteristics of the intervention and negatively impact the user's health status based on the temporal characteristics of the intervention. Specifically, the list determination unit analyzes the temporal characteristics of the medication. Drugs that show significant effectiveness for the user's symptoms within a specific timeframe are included in the intervention whitelist. Drugs that, although used, have a cumulative time exceeding a certain threshold but are ineffective or significantly less effective for the symptoms are included in the restricted intervention list. Drugs that are found to be ineffective for the user's symptoms and negatively impact their health status based on the temporal characteristics are included in the intervention contraindication list.

[0042] The feature library construction unit is used to perform beneficial and contraindicated intervention feature analysis based on the user's health characteristics, and, in conjunction with the intervention contraindication list, construct an intervention positive feature library and an intervention negative feature library. Specifically, the feature library construction unit performs beneficial and contraindicated intervention feature analysis based on the user's health characteristics, and, in conjunction with the intervention contraindication list, distinguishes between intervention positive features that can bring positive treatment effects and intervention negative features that may have adverse effects, thus constructing intervention positive feature libraries and intervention negative feature libraries, providing a basis for subsequent intervention recommendations and risk assessments.

[0043] The user intervention identification feature database construction unit is used to construct a user intervention identification feature database based on the intervention whitelist, restricted intervention list, intervention contraindication list, intervention positive feature database, and intervention negative feature database. Specifically, the user intervention identification feature database construction unit integrates information from the intervention whitelist, restricted intervention list, intervention contraindication list, intervention positive feature database, and intervention negative feature database. When a user buys or takes medication, whether by entering the requested drug name and ingredients or uploading photos, the unit can analyze the user's current health characteristics and intervention history to determine whether the proposed intervention is suitable for the user. If not, it provides corresponding guidance, such as suggesting changing the manufacturer or specifications, or recommending alternative interventions with similar effects but different ingredients. Since the user's intervention records are constantly changing, the user intervention identification feature database will also be dynamically updated to ensure that the provided intervention suggestions are always accurate and applicable.

[0044] In one possible implementation, the user medication identification feature library construction unit includes an intervention identification module generation subunit. This subunit is used to fit and construct an identification model based on the user intervention identification feature library, thereby generating an intervention identification module. Specifically, the intervention identification module generation subunit acquires data from the user intervention identification feature library, uses neural network algorithms to fit the data (e.g., it learns and models complex relationships in the data through multiple neuron layers, and uses techniques such as cross-validation to optimize the model's parameters, thereby improving the model's accuracy and generalization ability). Based on the algorithms and techniques, it constructs an identification model capable of accurately identifying user intervention patterns and needs. Based on the identification model, it generates an intervention identification module specifically used to determine the rationality of user interventions.

[0045] The user intervention guidance request acquisition subunit is used to acquire user intervention guidance requests, including drug ingredients and specifications, and manufacturer information. Specifically, the user drug guidance request acquisition subunit receives intervention guidance requests from users. When a user requests it, the subunit acquires detailed information provided by the user through a network interface or a dedicated application interface, including key elements such as the ingredients, specifications, and manufacturer of the drug of interest. Data verification technology is used to ensure the completeness and accuracy of the acquired information, providing a clear direction for subsequent analysis and guidance.

[0046] The request intervention analysis feature acquisition subunit is used to analyze drug characteristics and clinical data based on the drug ingredients, specifications, and manufacturer information to obtain the requested drug analysis features. Specifically, after receiving information such as drug ingredients, specifications, and manufacturer, the request drug analysis feature acquisition subunit analyzes the data, consults a drug feature database, and uses data mining techniques and statistical analysis methods to comprehensively analyze the drug's characteristics, efficacy, applicable symptoms, and possible side effects. For example, cluster analysis is used to group drugs with similar ingredients and effects together, principal component analysis is used to extract key drug features, a large amount of clinical data is referenced, and association rule mining techniques are used to discover the potential relationship between the drug and the treatment effect. Through a series of calculations and processing, analytical features that can accurately describe the characteristics and potential impact of the requested drug are extracted.

[0047] The user intervention guidance subunit is used to import the requested drug analysis features into the intervention identification module for user medication matching and identification, and output the requested medication identification result, which is used for user intervention guidance. Specifically, the user intervention guidance subunit imports the feature data obtained by the requested drug analysis feature acquisition subunit into the intervention identification module. The intervention identification module matches and identifies the features with the user's individual situation based on the previously constructed deep learning-based neural network model. For example, the neural network model will output the matching result through forward propagation calculation. After fast and accurate calculation and judgment, the requested intervention identification result is output. The result can clearly inform the user whether the requested drug is suitable for their own health condition and medication history, providing scientific and accurate guidance for the user's medication decision.

[0048] The management and adjustment module 50 is used to open a verification feedback interface based on the user health intervention feature map. This interface connects to a user terminal to annotate and provide feedback on features in the user health intervention feature map. Based on the feedback data obtained through the verification feedback interface, the management and adjustment module 50 manages and adjusts the user health intervention feature map. Specifically, the management and adjustment module 50 opens a verification feedback interface based on the user health intervention feature map. This interface is opened to establish a direct connection with the user terminal, enabling users to participate in the evaluation and improvement of their own medical feature database. Users access this verification feedback interface through terminal devices, such as mobile applications or web pages. In the interface provided, users can annotate and provide feedback on various features presented in the user health intervention feature map in detail. For example, if a user believes that a certain medication feature record is inaccurate, or that the description of a health feature does not match the actual situation, they can make clear annotations and explanations in the corresponding locations. The management and adjustment module 50 collects user feedback data obtained through the verification feedback interface in real time. This data includes user feedback on the medical feature database. Based on feedback data, the management and adjustment module 50 manages and adjusts the user health intervention feature map by collecting opinions, suggestions, and corrections on various features. This includes correcting inaccurate data, supplementing missing information, and re-evaluating and setting data classifications and weights based on user feedback. For example, if multiple users report discrepancies between the effects of a certain drug and the records in the feature database, the management and adjustment module 50 will re-examine the relevant data, potentially conducting further verification and confirmation with medical professionals. Then, it will update and adjust the feature records for that drug. Through this iterative process, the user health intervention feature map is continuously optimized and improved based on user feedback, making it more accurate and better suited to users' actual situations, thereby providing users with higher-quality and more personalized medical services and advice.

[0049] In the above text, refer to Figure 1 A user health intervention feature management system based on multi-source data according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 This invention describes a user health intervention feature management method based on multi-source data according to embodiments of the present invention.

[0050] User health intervention feature management methods based on multi-source data, such as Figure 2 As shown, the method includes:

[0051] User identity information is verified through an authorized channel. Based on the verified user identity information, a health intervention management platform is connected to obtain a user health intervention time-series dataset, which includes platform intervention records, non-platform intervention records, and self-managed health behavior records. Unstructured health records provided by the user are parsed to generate standardized data, and a supplementary extended database is constructed. Non-health intervention management platform intervention data is aligned using a knowledge graph. The time-series dataset, unstructured health records, and non-health intervention management platform intervention data are fused to form a user-coordinated dataset. User health features and intervention features are analyzed based on the user-coordinated dataset to obtain a user health intervention feature map. A verification feedback interface is opened based on the user health intervention feature map. This interface connects to the user terminal to annotate and provide feedback on the features in the user health intervention feature map. The user health intervention feature map is managed and adjusted based on the feedback data obtained from the verification feedback interface.

[0052] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: setting authentication rules, the authentication rules including authentication type, number of authentications, and authentication order, the authentication type including ID card, password, and QR code; extracting authentication features according to the authentication rules, establishing a temporal relationship between the authentication type and the authentication order, and fitting and constructing a user authentication module; verifying sequentially according to the temporal relationship through the user authentication module to obtain authentication results; when the authentication result is successful, generating authorization information to establish a communication connection with the health intervention management platform.

[0053] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: establishing a data addition interface for scanning or entering unstructured records; identifying the added data using recognition technology to determine multi-dimensional identification information of the added data, including intervention plan, implementation time, and physiological indicators; performing similarity calculations between the multi-dimensional identification information and the user health intervention time-series dataset to determine newly added data; and identifying the type of the added data and aligning non-medical management platform intervention data using a knowledge graph.

[0054] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: cleaning the user data set and standardizing it; extracting user health data and intervention record data from the preprocessed user data set, and constructing time-series relationships and extracting features from the user health data and intervention record data respectively, including frequency, interval, health status or intervention change features; quantifying and encoding the extracted features to construct feature vectors; performing feature clustering based on the feature vectors to identify user health feature categories and intervention feature categories; using association mining algorithms to analyze the correlation between user health features and intervention features; establishing a mapping association between the feature vectors and the user health feature categories and medication feature categories, and fitting the correlation to construct the user health intervention feature map.

[0055] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: constructing an original extracted feature matrix, learning the nonlinear interaction between features through a multilayer perceptron, and outputting new higher-order features; concatenating the output higher-order features with the original features to obtain an added feature matrix; defining a regularized loss function, training a regression model and obtaining the optimal weights; and performing feature selection on the added feature matrix based on the regression model, filtering features with non-zero weights.

[0056] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: extracting user health features and intervention features according to the user medical feature database, and establishing a temporal relationship between the medication features and the user health features; performing temporal feature analysis based on the medication features to determine an intervention whitelist, a restricted intervention list, and an intervention contraindication list, wherein the medication whitelist is a list of interventions effective for user symptoms in the temporal features of intervention, the restricted intervention list is a list of drugs whose cumulative intervention time exceeds a threshold and are ineffective or significantly weakened in treating user symptoms in the temporal features of intervention, and the intervention contraindication list is a list of interventions that are ineffective for user symptoms and affect user health status in the temporal features of intervention; performing beneficial intervention and contraindicated intervention feature analysis based on the user health features, and constructing a positive intervention feature database and a negative intervention feature database in combination with the intervention contraindication list; and constructing a user intervention identification feature database based on the intervention whitelist, restricted intervention list, intervention contraindication database, positive intervention feature database, and negative intervention feature database.

[0057] In one possible implementation, the user health intervention feature management method based on multi-source data further includes: fitting and constructing an identification model based on the user intervention identification feature library to generate an intervention identification module; obtaining user intervention guidance requests, including drug ingredients and specifications, and manufacturer information; performing drug feature and clinical data analysis based on the drug ingredients, specifications, and manufacturer information to obtain requested drug analysis features; importing the requested drug analysis features into the intervention identification module for user medication matching and identification, and outputting the requested medication identification result, which is used for user intervention guidance.

[0058] The user health intervention feature management system based on multi-source data provided in the embodiments of the present invention can execute the user health intervention feature management method based on multi-source data provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0059] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0060] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A user health intervention feature management system based on multi-source data, characterized in that, The user health intervention feature management system based on multi-source data comprises: A medical data acquisition module, configured to verify user identity information through an authorized channel, connect a health intervention management platform based on the verified user identity information, and acquire a user health intervention time series data set, wherein the time series data set comprises platform intervention records, non-platform intervention records, and autonomous health behavior records; A supplementary expansion database construction module, configured to parse unstructured health records provided by a user, generate standardized data, construct a supplementary expansion database, and align non-health intervention management platform intervention data in the database through a knowledge graph; A comprehensive data set assembly module, configured to fuse the time series data set, the unstructured health records, and non-health intervention management platform intervention data to assemble a user comprehensive data set; A medical feature library acquisition module, configured to analyze user health features and intervention features based on the user comprehensive data set, and obtain a user health intervention feature graph; A management adjustment module, configured to open a verification feedback interface based on the user health intervention feature graph, the verification feedback interface being configured to connect a user terminal to label and feed back features in the user health intervention feature graph, and adjust the user health intervention feature graph based on feedback data acquired through the verification feedback interface; The medical feature library acquisition module comprises: A standardization processing unit, configured to clean and standardize the user comprehensive data set; A data set extraction unit, configured to extract user health data and intervention record data from the preprocessed user comprehensive data set, construct time series relationships and extract features from the user health data and intervention record data, and the extracted features include frequency, interval, health status, and intervention change features; A feature vector construction unit, configured to quantize and encode the extracted features to construct feature vectors; A feature clustering unit, configured to cluster features based on the feature vectors to identify user health feature categories and intervention feature categories; A correlation analysis unit, configured to analyze the correlation between user health features and intervention features by using an association mining algorithm; A medical feature library construction unit, configured to map and correlate the feature vectors with the user health feature categories and intervention feature categories, fit the correlation, and construct the user health intervention feature graph; The medical feature library construction unit comprises: A high-order feature output subunit, configured to construct an original extracted feature matrix, learn the nonlinear interaction between features through a multi-layer perception machine, and output new high-order features; A feature splicing subunit, configured to splice the output high-order features with the original features to obtain an increased feature matrix. The regularization loss function definition subunit is configured to define a regularization loss function, train a regression model, and obtain optimal weights. The feature selection subunit is configured to perform feature selection on the increased feature matrix based on the regression model, and filter features with non-zero weights.

2. The multi-source data based user health intervention feature management system as claimed in claim 1, wherein, The medical data acquisition module includes: The identity verification rule setting unit is configured to set identity verification rules, including identity verification types, verification quantities, and verification sequences, and the identity verification types include identity cards, passwords, and two-dimensional codes. The verification module construction unit is configured to extract verification features according to the identity verification rules, establish a time sequence relationship between the identity verification types and the verification sequences, and fit and construct a user verification module. The verification result acquisition unit is configured to sequentially verify through the user verification module according to the time sequence relationship and obtain a verification result. The communication connection unit is configured to generate authorization information when the verification result is verified, and establish a communication connection of the health intervention management platform.

3. The multi-source data based user health intervention feature management system as claimed in claim 1, wherein, The supplementary expansion database construction module includes: The adding interface establishment unit is configured to establish a data adding interface for scanning or inputting user unstructured records. The added data recognition unit is configured to recognize added data through recognition technology, determine multi-dimensional recognition information of the added data, including intervention schemes, implementation times, and physiological indicators. The data duplication checking unit is configured to perform similarity operation on the multi-dimensional recognition information and a user health intervention time series data set to determine newly added data. The type recognition unit is configured to recognize the types of the added data, and align non-medical management platform intervention data thereof through a knowledge graph.

4. The multi-source data based user health intervention feature management system as claimed in claim 1, wherein, The medical feature library acquisition module includes: The time sequence relationship establishment unit is configured to extract user health features and intervention features from the user medical feature library, and establish a time sequence relationship between the medication features and the user health features. The list determination unit is configured to perform time sequence feature analysis according to the medication features, determine an intervention white list, a restricted intervention list, and an intervention contraindication list, the intervention white list is an intervention list that is effective for user symptoms in the intervention time sequence features, the restricted intervention list is a drug list that has an intervention cumulative time exceeding a threshold value and is ineffective or has significantly weakened effect on user symptoms in the intervention time sequence features, and the intervention contraindication list is an intervention list that is ineffective for user symptoms and affects user health status in the time sequence features. The feature library construction unit is configured to perform beneficial intervention and contraindication intervention feature analysis according to the user health features, and construct a positive intervention feature library and a negative intervention feature library in combination with the intervention contraindication list. The user intervention identification feature library construction unit is configured to construct a user intervention identification feature library according to the intervention whitelist, the intervention restriction blacklist, the intervention taboo blacklist, the intervention positive feature library, and the intervention negative feature library.

5. The multi-source data based user health intervention feature management system of claim 4, wherein, The user medication identification feature library construction unit includes: The intervention identification module generation subunit is configured to fit and construct an identification model based on the user intervention identification feature library, and generate an intervention identification module. The user intervention guidance request acquisition subunit is configured to acquire a user intervention guidance request, which includes information about drug components and specifications, and manufacturers. The request intervention analysis feature acquisition subunit is configured to perform drug feature and clinical data analysis based on the information about drug components and specifications, and manufacturers, and obtain request drug analysis features. The user intervention guidance subunit is configured to import the request drug analysis features into the intervention identification module for user medication matching and identification, and output a request medication identification result, which is used for user intervention guidance.

6. A method for managing features of a user health intervention based on multi-source data, the method comprising: The method is applied to the user health intervention feature management system based on multi-source data according to any one of claims 1-5, and the method includes: Verify user identity information through an authorized channel, connect a health intervention management platform based on the verified user identity information, acquire a user health intervention time series data set, and the time series data set includes platform intervention records, non-platform intervention records, and autonomous health behavior records; Parse unstructured health records provided by a user, generate standardized data, construct a supplementary extension database, and align non-health intervention management platform intervention data through a knowledge graph; Fuse the time series data set, the unstructured health records, and the non-health intervention management platform intervention data to form a user overall data set; Analyze user health features and intervention features based on the user overall data set to obtain a user health intervention feature map; Open a verification feedback interface according to the user health intervention feature map, the verification feedback interface is configured to connect a user terminal to mark and feed back features in the user health intervention feature map, and manage and adjust the user health intervention feature map based on feedback data acquired through the verification feedback interface.

Images