Traditional Chinese medicine health preserving intelligent pushing system based on knowledge graph

Abstract

The invention discloses a traditional Chinese medicine health preserving intelligent pushing system based on a knowledge graph, and the system comprises a user data collection module which is used for collecting basic information, health conditions and living habits of a user, and behavior data in the system; the data analysis module is used for deeply analyzing the user data and identifying the interest preference and demand characteristics of the user and the association relationship between the user behavior and the health care knowledge; the knowledge graph construction module is used for constructing a knowledge graph based on professional knowledge in the field of traditional Chinese medicine health care; the intelligent recommendation module is used for generating a customized traditional Chinese medicine health care knowledge and service recommendation list for the user according to the user portrait and the knowledge graph; and the interactive interface module is used for displaying the recommended content and performing interactive operation between the user and the system. According to the system, through accurate recommendation, the user experience is improved, the problem that traditional Chinese medicine life cultivation and health preservation knowledge spreading is lack of pertinence and individuation is effectively solved, and meanwhile the diversified requirements of users are met.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent recommendation systems, and in particular to a knowledge graph-based intelligent push system for traditional Chinese medicine health care. Background Art

[0002] As people's health awareness increases, the demand for TCM health care knowledge and services is growing rapidly. However, the existing TCM health care information dissemination methods have many shortcomings. First of all, the dissemination of this information is often lacking in pertinence and personalization, and it is difficult for users to quickly obtain content that meets their needs and interests from massive amounts of information. Traditional information recommendation systems mostly rely on simple classification or keyword matching, which makes it difficult to accurately understand users' deep-seated needs and interest preferences, resulting in unsatisfactory recommendation results and poor user experience.

[0003] Secondly, TCM health care knowledge is highly professional and diverse, covering TCM theory, diet therapy, exercise health care, meridian acupoints and other aspects. These complex knowledge systems make it difficult for users to obtain and understand them. How to systematically organize this knowledge and accurately recommend it to users in need is an important technical problem facing the current field of TCM health care information dissemination. Summary of the invention

[0004] In order to solve the problems raised in the above background technology, the present invention provides a Chinese medicine health care intelligent push system based on knowledge graph, which can more accurately understand and match user needs with health care knowledge content by integrating multi-source data, introducing knowledge graph technology and dynamically adjusting recommendation strategies in real time.

[0005] In order to solve the above technical problems, the technical solution proposed by the present invention is:

[0006] A TCM health care intelligent push system based on knowledge graph, characterized by comprising: a user data collection module, a data analysis module, a knowledge graph construction module, an intelligent recommendation module and an interactive interface module;

[0007] The user data collection module is used to collect the user's basic information, health status, living habits and behavior data in the system to build a user portrait;

[0008] The data analysis module conducts in-depth analysis of user data to identify user interest preferences, demand characteristics, and the relationship between user behavior and health care knowledge, providing a basis for subsequent personalized recommendations;

[0009] The knowledge graph construction module constructs a knowledge graph containing entities such as TCM theories, symptoms, health preservation methods, food materials, medicinal materials, and their relationships based on professional knowledge in the field of TCM health care;

[0010] The intelligent recommendation module generates a customized list of TCM health care knowledge and service recommendations for users based on user portraits and knowledge graphs using a personalized recommendation algorithm. The recommendation algorithm comprehensively considers the user's interest preferences, health status, current season, and regional factors.

[0011] The interactive interface module is used to display recommended content, receive user feedback, and allow the user to interact with the system.

[0012] Preferably, the user data collection module further comprises:

[0013] The basic information collection unit is used to collect the user's basic information, including age, gender, and occupation, through the user registration and information completion links set up in the system;

[0014] A health status collection unit, used to collect the user's health status information, including past medical history and current symptoms;

[0015] Lifestyle collection unit, used to collect user's lifestyle data, including dietary preferences, exercise habits, and sleep conditions;

[0016] The behavior data recording unit is used to record the user's interactive behavior data in the system, including browsing history, search keywords, likes, and comments.

[0017] Preferably, the user data collection module obtains basic health information and medical history data by cooperating with medical institutions and health management platforms, and sets up user registration and information completion links in the system to guide users to fill in information, while tracking and recording user behavior data in real time.

[0018] Preferably, the data analysis module further comprises:

[0019] The user portrait construction unit uses machine learning algorithms, including cluster analysis, classification algorithms, and association rule mining algorithms, to conduct in-depth analysis of user data;

[0020] Demand feature extraction unit, which extracts key demand features from the user's health status and living habits data;

[0021] The association rule mining unit mines the potential correlation between user behavior and health care knowledge. By analyzing a large amount of user behavior data, including browsing records and search keywords, it discovers the correlation between user behavior patterns in different scenarios and health care knowledge, providing a basis for subsequent personalized recommendations.

[0022] Preferably, the user portrait construction unit first performs a comprehensive processing of basic information, health status, living habits and behavioral data, classifies users with similar characteristics into one category through a clustering algorithm, and then classifies the different attributes of the users through a decision tree algorithm. Finally, an association rule mining algorithm is used to mine the potential correlation between the data, thereby constructing a comprehensive user portrait, which contains the user's dimensional information, including basic information, health status, interest preferences, and behavioral characteristics.

[0023] Preferably, the knowledge graph construction module further includes:

[0024] The knowledge sorting and classification unit organizes TCM experts and knowledge engineers to systematically sort out and classify TCM health care knowledge, clarifying that the entities of the knowledge graph include TCM theory, symptoms, health care methods, food ingredients, and medicinal materials, and determining the relationship types between entities, including the correspondence between TCM theory and symptoms, and the matching relationship between health care methods and food ingredients. Through expert experience and knowledge system construction, the professionalism and rationality of the knowledge graph are ensured;

[0025] The data collection and annotation unit collects data from multiple channels, including traditional Chinese medicine classics, academic literature, and clinical experience, and uses natural language processing techniques including named entity recognition and relationship extraction to extract and annotate text data to extract entity and relationship information;

[0026] The graph database storage unit uses a graph database to store the constructed knowledge graph to support efficient query and reasoning operations. The graph database can better handle the complex relationships in the knowledge graph, so that relevant TCM health care knowledge can be quickly retrieved and associated during the recommendation process, improving the system's response speed and recommendation accuracy.

[0027] Preferably, the intelligent recommendation module further comprises:

[0028] The content-based recommendation unit selects matching TCM health care knowledge and services from the knowledge graph based on the characteristics of the user portrait, including interest preferences and health status;

[0029] The collaborative filtering recommendation unit analyzes the behavioral data of user groups with similar interests and needs through user clustering and behavior pattern mining technologies. It first clusters users to find user groups with similar interests and needs, and then mines the behavior patterns of the groups. The collaborative filtering algorithm discovers potential connections between users and recommends content that target users are generally interested in, thereby improving the diversity and accuracy of recommendations.

[0030] The dynamic weight adjustment unit dynamically adjusts the weights of different factors according to the user's real-time behavior and feedback information to ensure the accuracy and practicality of the recommendation. At the same time, it considers the impact of external factors including season and region on the recommendation results, and adjusts the weights of various factors in the recommendation algorithm in a timely manner according to the user's real-time browsing and search behavior and feedback on the recommended content, so that the recommendation results are more in line with the user's current needs and interests.

[0031] Preferably, the interactive interface module further comprises:

[0032] The homepage recommendation unit displays a list of health care knowledge and service recommendations customized for users based on user portraits and real-time data, presented in a graphic and text format to attract users' attention;

[0033] Search function unit, users can enter keywords for precise search according to their specific needs. Keywords include disease names and health methods. The system matches and displays relevant search results in the knowledge graph based on the keywords entered by users, and provides multiple filtering conditions including knowledge type, applicable population, and season, and presents accurate search results to users according to the filtering conditions;

[0034] The personal center unit is used by users to view and manage their own information such as basic information, health status, living habits, etc., set preferences, view recommendation history and feedback records, so as to facilitate users to manage their own health learning process. Users can update their information at any time in the personal center, adjust health preference settings, view the content recommended by the system in the past and their feedback on these recommendations, so as to better understand the changes in their health needs and the effects of system recommendations, and realize personalized health management;

[0035] The knowledge classification browsing unit displays knowledge according to different fields of TCM health care, including diet therapy, exercise, meridian health care, and TCM theory, making it easier for users to find content of interest. Users can browse by selecting the corresponding category according to their field of interest, thus improving the efficiency and accuracy of users' knowledge search;

[0036] Preferably, the TCM health care intelligent push system based on knowledge graph supports access from multiple terminal devices, including mobile phones, computers, and tablets, providing users with convenient and efficient TCM health care knowledge intelligent recommendation services.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention not only considers the user's basic information and behavior data, but also deeply mines multi-source data such as the user's health status and living habits to build a more comprehensive and accurate user portrait. This comprehensive analysis of multi-dimensional data enables the system to have a deeper understanding of the user's potential needs and interest preferences, thereby providing users with more accurate recommendation services.

[0039] 2. The present invention introduces knowledge graph technology to structure and associate the complex knowledge in the field of TCM health care. The system can understand the deep relationship between user needs and health care knowledge based on the semantic information of the knowledge graph, and achieve smarter and more accurate recommendations. This semantic understanding and matching capability enables the system to provide users with comprehensive health care solutions to meet the diverse needs of users.

[0040] 3. The present invention can timely track changes in user behavior and fluctuations in health status, and adjust recommendation strategies in real time. At the same time, the system will dynamically adjust the recommended content according to external factors such as seasonal changes and epidemic disease trends, ensuring that users can always obtain health care recommendations that best meet current actual conditions. This dynamic adjustment capability improves the accuracy and practicality of recommendations and enhances users' trust and reliance on the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 is a system architecture diagram of the present invention;

[0042] Figure 2 It is a system workflow diagram of the present invention;

[0043] Figure 3 This is the system knowledge graph construction process of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] Example 1

[0046] refer to Figure 1 to Figure 2 ,The workflow of a TCM health intelligent push system based on knowledge graph is as follows:

[0047] First, the system starts working through the user data collection module. This module includes a basic information collection unit, a health status collection unit, a living habit collection unit, and a behavior data recording unit. When a user uses the system for the first time or updates personal information, he or she needs to register and complete personal information, such as basic information such as age, gender, and occupation. At the same time, the system obtains the user's health status information, including past medical history, current symptoms, etc., by cooperating with medical institutions and health management platforms. In addition, the system will also collect the user's living habit data, such as dietary preferences, exercise habits, sleep conditions, etc., and record the user's behavior data in the system, such as browsing history, search keywords, likes, comments, etc. These information together form the basis of the user portrait.

[0048] Next, the data analysis module conducts an in-depth analysis of the user data. This module includes a user portrait construction unit, a demand feature extraction unit, and an association rule mining unit. The user portrait construction unit uses machine learning algorithms, such as cluster analysis, classification algorithms, and association rule mining algorithms, to process user data and construct a user portrait containing multi-dimensional information such as user basic information, health status, interest preferences, and behavioral characteristics. The demand feature extraction unit extracts key demand features from the user's health status and living habits data. The association rule mining unit mines the potential correlation between user behavior and health care knowledge to provide a basis for subsequent personalized recommendations.

[0049] In the knowledge graph construction module, the system builds a knowledge graph containing entities such as TCM theory, symptoms, health methods, food ingredients, medicinal materials and their relationships based on professional knowledge in the field of TCM health care. This module includes a knowledge combing and classification unit, a data collection and annotation unit and a graph database storage unit. The knowledge combing and classification unit organizes TCM experts and knowledge engineers to systematically comb and classify TCM health care knowledge, and clarify the entity and relationship types of the knowledge graph. The data collection and annotation unit collects data from multiple channels such as TCM classics, academic literature, clinical experience, etc., and uses natural language processing technology to extract and annotate the data. The graph database storage unit uses a graph database to store the constructed knowledge graph to support efficient query and reasoning operations.

[0050] The intelligent recommendation module uses a personalized recommendation algorithm to generate a customized list of TCM health care knowledge and service recommendations for users based on user portraits and knowledge graphs. This module includes a content-based recommendation unit, a collaborative filtering recommendation unit, and a dynamic weight adjustment unit. The content-based recommendation unit selects matching TCM health care knowledge and services from the knowledge graph based on the features in the user portrait. The collaborative filtering recommendation unit analyzes the behavioral data of user groups with similar interests and needs through user clustering and behavioral pattern mining technology, and recommends content of general interest to target users. The dynamic weight adjustment unit dynamically adjusts the weights of different factors based on the user's real-time behavior and feedback information to ensure the accuracy and practicality of the recommendation. At the same time, the module will also consider the impact of external factors such as season and region on the recommendation results, and adjust the recommendation strategy in a timely manner.

[0051] Finally, the interactive interface module displays the recommended content to the user and receives user feedback. This module includes a homepage recommendation unit, a search function unit, a personal center unit, and a knowledge classification browsing unit. The homepage recommendation unit displays a list of health care knowledge and service recommendations customized for users based on user portraits and real-time data in the form of pictures and texts. The search function unit allows users to enter keywords for precise searches based on their specific needs, and provides a variety of filtering conditions for users to choose from. The personal center unit is used for users to view and manage their own information, set preferences, view recommendation history and feedback records, etc. The knowledge classification browsing unit displays knowledge in categories according to different fields of traditional Chinese medicine health care, making it easier for users to find content of interest.

[0052] Example 2

[0053] refer to Figure 3 The knowledge graph construction module is based on the professional knowledge in the field of TCM health care, and constructs a knowledge graph containing entities such as TCM theory, symptoms, health care methods, food ingredients, medicinal materials and their relationships. The following is the complete process of knowledge graph construction:

[0054] First, organize TCM experts and knowledge engineers to sort out and classify the collected knowledge, determine the entities of the knowledge graph, and determine the relationship types between entities. Entities include TCM theory, symptoms, health preservation methods, food materials, and medicinal materials. Specifically, TCM theory includes the Yin-Yang balance theory, meridian theory, etc., symptoms include colds, hypertension, cervical spondylosis, etc., health preservation methods include Tai Chi, diet therapy, moxibustion, etc., food materials include celery, wolfberry, yam, etc., and medicinal materials include ginseng, astragalus, angelica, etc.

[0055] Specifically, the named entity recognition method in natural language processing technology is used to perform entity recognition on text data. The conditional random field model is used for named entity recognition. Suppose the text data is T = {t1, t2, ..., t n}, the label set is L = {l1,l2,...,l m}, the goal of the conditional random field model is to find a label sequence Y = {y1,y2,...,y n}, so that the probability P(Y|T) is maximized. By training the conditional random field model, it can accurately identify entities such as diseases, ingredients, and medicinal materials in the text.

[0056] Specifically, a rule-based and machine learning method is used to extract relations.

[0057] Specifically, rule-based relationship extraction formulates rules based on specific sentence structures and semantic patterns in TCM knowledge. For example, for a sentence pattern such as "a certain disease is suitable for a certain health-preserving method", the relationship between the disease and the health-preserving method is extracted. For a sentence pattern such as "a certain medicinal material has a therapeutic effect on a certain disease", the relationship between the medicinal material and the disease is extracted.

[0058] For the text "For a cold, it is advisable to take the health-preserving approach of resting and drinking plenty of water", the relationship between "cold" and "rest and drink plenty of water" is extracted as "treatment advice" according to the rules.

[0059] Specifically, the relationship extraction based on machine learning uses a convolutional neural network for relationship extraction training. The labeled text data is used as a training set, where the input is the entity pairs in the text data and their context information, and the output is the relationship between the entity pairs. For example, for the input "celery" and "diet therapy" and their context information, the convolutional neural network model can output the relationship between them as "ingredients are used for diet therapy" after training.

[0060] Specifically, the graph database selects Neo4j graph database to store the constructed knowledge graph. Neo4j graph database has a good ability to handle complex relationships and can efficiently support the query and reasoning operations of the knowledge graph. In the graph database, the entities of the knowledge graph are taken as nodes, and the relationships between entities are stored as edges. For example, the "cold" symptom node is connected to the "rest and drink plenty of water" health method node by an edge, and the attribute of the edge is "treatment advice". Similarly, the "celery" ingredient node is connected to the "diet therapy" health method node by an edge, and the attribute of the edge is "ingredients used for diet therapy".

[0061] Specifically, when new academic research results in the field of TCM health care are published, or TCM doctors have new experiences and discoveries in clinical practice, this new information needs to be updated to the knowledge graph.

[0062] Specifically, by analyzing user feedback and behavior data in the system, if we find that users' needs for certain knowledge are not met, or if we find that users have new related understandings of certain knowledge, we need to update the knowledge graph.

[0063] Specifically, for new knowledge information, first determine the entities and relationships involved. For example, for a new ingredient, determine its relationship with other entities including symptoms, health methods, etc. Then, according to the method of knowledge graph construction, add the new entity as a node and the new relationship as an edge to the graph database.

[0064] Specifically, after adding new entities and relationships, the knowledge graph needs to be retrained and optimized. For example, for the relationship extraction model based on machine learning, the model needs to be retrained with updated annotated data to improve the accuracy and generalization ability of the model. The attributes of the nodes and edges in the knowledge graph need to be adjusted and optimized based on the new knowledge to ensure the accuracy and consistency of the knowledge graph.

[0065] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A TCM health care intelligent push system based on knowledge graph, characterized in that: include: User data collection module, data analysis module, knowledge graph construction module, intelligent recommendation module and interactive interface module; The user data collection module is used to collect the user's basic information, health status, living habits and behavior data in the system to build a user portrait; The data analysis module performs in-depth analysis on user data to identify user interest preferences, demand characteristics, and the relationship between user behavior and health care knowledge; The knowledge graph construction module constructs a knowledge graph containing entities such as TCM theories, symptoms, health preservation methods, food materials, medicinal materials, and their relationships based on professional knowledge in the field of TCM health care; The intelligent recommendation module generates a customized list of TCM health care knowledge and service recommendations for users based on user portraits and knowledge graphs using a personalized recommendation algorithm. The recommendation algorithm comprehensively considers the user's interest preferences, health status, current season, and regional factors. The interactive interface module is used to display recommended content, receive user feedback, and allow the user to interact with the system.

2. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The user data collection module further includes: Basic information collection unit, used to collect basic information of users, including age, gender, and occupation; A health status collection unit, used to collect the user's health status information, including past medical history and current symptoms; Lifestyle collection unit, used to collect user's lifestyle data, including dietary preferences, exercise habits, and sleep conditions; The behavior data recording unit is used to record the user's interactive behavior data in the system, including browsing history, search keywords, likes, and comments.

3. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The user data collection module obtains basic health information and medical history data by cooperating with medical institutions and health management platforms, and sets up user registration and information completion links in the system to guide users to fill in information, while tracking and recording user behavior data in real time.

4. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The data analysis module further comprises: The user portrait construction unit uses machine learning algorithms, including cluster analysis, classification algorithms, and association rule mining algorithms, to conduct in-depth analysis of user data; Demand feature extraction unit, which extracts key demand features from the user's health status and living habits data; The association rule mining unit mines the potential correlation between user behavior and health care knowledge. By analyzing a large amount of user behavior data, including browsing records and search keywords, it discovers the correlation between user behavior patterns in different scenarios and health care knowledge.

5. According to claim 4, a TCM health care intelligent push system based on knowledge graph is characterized in that: The user portrait construction unit first comprehensively processes the basic information, health status, living habits and behavioral data, classifies users with similar characteristics into one category through a clustering algorithm, and then classifies the different attributes of the users through a decision tree algorithm. Finally, an association rule mining algorithm is used to mine the potential correlation between the data, thereby constructing a comprehensive user portrait, which contains the user's dimensional information, including basic information, health status, interest preferences, and behavioral characteristics.

6. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The knowledge graph construction module further includes: The knowledge sorting and classification unit organizes TCM experts and knowledge engineers to systematically sort out and classify TCM health care knowledge, clarifying that the entities of the knowledge graph include TCM theory, symptoms, health care methods, food ingredients, and medicinal materials, and determining the relationship types between entities, including the correspondence between TCM theory and symptoms, and the matching relationship between health care methods and food ingredients; The data collection and annotation unit collects data from multiple channels, including traditional Chinese medicine classics, academic literature, and clinical experience, and uses natural language processing techniques including named entity recognition and relationship extraction to extract and annotate text data to extract entity and relationship information; The graph database storage unit uses the graph database to store the constructed knowledge graph.

7. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The intelligent recommendation module further includes: The content-based recommendation unit selects matching TCM health care knowledge and services from the knowledge graph based on the characteristics of the user portrait, including interest preferences and health status; The collaborative filtering recommendation unit analyzes the behavioral data of user groups with similar interests and needs through user clustering and behavioral pattern mining technologies. It first clusters users to find user groups with similar interests and needs, and then mines the behavioral patterns of the groups. The collaborative filtering algorithm discovers potential connections between users and recommends content that they are generally interested in to target users. The dynamic weight adjustment unit dynamically adjusts the weights of different factors according to the user's real-time behavior and feedback information to ensure the accuracy and practicality of the recommendation. At the same time, it considers the impact of external factors including season and region on the recommendation results, and adjusts the weights of various factors in the recommendation algorithm in a timely manner according to the user's real-time browsing and search behavior and feedback on the recommended content, so that the recommendation results are more in line with the user's current needs and interests.

8. According to claim 1, a TCM health care intelligent push system based on knowledge graph is characterized in that: The interactive interface module further comprises: The homepage recommendation unit displays a list of health care knowledge and service recommendations customized for users based on user portraits and real-time data; Search function unit, users can enter keywords for precise search according to their specific needs. Keywords include disease names and health methods. The system matches and displays relevant search results in the knowledge graph based on the keywords entered by users, and provides multiple filtering conditions including knowledge type, applicable population, and season, and presents accurate search results to users according to the filtering conditions; The personal center unit is used by users to view and manage their own information, set preferences, view recommendation history and feedback records. Users can update their information at any time in the personal center, adjust health preference settings, view the content recommended by the system in the past and their feedback on these recommendations; The knowledge classification browsing unit displays knowledge according to different fields of TCM health care, including diet therapy, exercise, meridian health care, and TCM theory. Users can choose the corresponding category to browse according to their areas of interest.

9. According to claims 1 to 8, a TCM health care intelligent push system based on knowledge graph is characterized in that: The intelligent push system for traditional Chinese medicine health care based on knowledge graph supports access from multiple terminal devices, including mobile phones, computers, and tablets.

Images