A Personalized Recommendation Method and System for Traditional Chinese Medicine Wellness Programs Based on Big Data

Abstract

This invention provides a method and system for personalized recommendation of TCM health preservation plans based on big data. By constructing a primary health preservation interest graph for target users, it accurately captures users' interests and needs in the field of TCM health preservation knowledge. Utilizing knowledge reinforcement coefficients and interest mapping factors, it deeply mines and personally matches knowledge nodes in the TCM health preservation knowledge domain, generating an optimized health preservation interest graph. This not only improves the accuracy and personalization of TCM health preservation plan recommendations but also continuously iterates and optimizes the recommendation results through a cyclical optimization mechanism, effectively meeting the TCM health preservation needs of target users and improving the practicality and user satisfaction of the health preservation plans.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and more specifically, to a method and system for personalized recommendation of TCM health preservation plans based on big data. Background Technology

[0002] With the rapid development of big data technology and people's increasing emphasis on health and wellness, the field of traditional Chinese medicine (TCM) health preservation is gradually integrating with big data technology, aiming to provide individuals with more accurate and personalized health preservation plans through data analysis. However, existing TCM health preservation recommendation systems mostly rely on traditional static recommendation methods or simple data analysis, making it difficult to comprehensively capture users' dynamic health needs and interests.

[0003] Traditional Chinese medicine (TCM) health preservation recommendations often rely on fixed rules or expert experience, lacking in-depth exploration and dynamic adjustment of users' personalized needs. Furthermore, existing data analysis methods often struggle to effectively extract and utilize the relationships between knowledge nodes when dealing with the complex knowledge system of TCM health preservation, resulting in inaccurate and incomplete recommendation results. Summary of the Invention

[0004] In view of the aforementioned problems, and in conjunction with the first aspect of the present invention, embodiments of the present invention provide a method for personalized recommendation of TCM health preservation plans based on big data, the method comprising:

[0005] Based on multiple health feedback data of the target user and the tracking parameters corresponding to each health feedback data, a first health interest graph in the TCM health knowledge field of the target user is constructed; the first health interest graph includes the first correlation degree between each knowledge node in the TCM health knowledge field and the TCM health knowledge profile of the target user;

[0006] Based on the first relevance of each knowledge node and the predefined prior attention trend, the knowledge reinforcement coefficient and interest mapping factor of the corresponding knowledge node are determined respectively;

[0007] Based on the knowledge enhancement coefficient, all knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing to generate multiple extended mining nodes. Based on the interest mapping factor, interest mapping is performed on the multiple extended mining nodes to generate an interest mapping vector for each extended mining node.

[0008] Based on the interest mapping vector of the extended mining node, the first health care interest graph of the target user is iteratively optimized to generate an optimized health care interest graph in the field of traditional Chinese medicine health care knowledge of the target user.

[0009] Based on the optimized health and wellness interest graph, a traditional Chinese medicine (TCM) health and wellness plan is constructed for the target user, and a TCM health and wellness plan recommended to the target user is generated.

[0010] In one possible implementation of the first aspect, the step of constructing a first health-preservation interest map in the field of traditional Chinese medicine health-preservation knowledge of the target user based on multiple health feedback data of the target user and tracking parameters corresponding to each health feedback data includes:

[0011] Extract the health preservation scope information and health preservation monitoring parameters of the TCM health preservation knowledge field from the tracking parameters;

[0012] The target user's area of ​​TCM health preservation knowledge is determined based on the aforementioned health preservation scope information;

[0013] Based on the aforementioned health monitoring parameters, determine the attribute information of each knowledge node in the domain of traditional Chinese medicine health preservation knowledge;

[0014] The attribute information and the multiple health feedback data are loaded into the TCM health preservation knowledge association model to generate the first association degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user;

[0015] By utilizing the first correlation degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user, a first health preservation interest map of the TCM health preservation knowledge domain in which the target user resides is constructed.

[0016] In one possible implementation of the first aspect, determining the knowledge enhancement coefficient and interest mapping factor of the corresponding knowledge node based on the first relevance of each knowledge node and a predefined prior attention trend includes:

[0017] Based on the first degree of relevance of each knowledge node, determine the knowledge reinforcement coefficient of the corresponding knowledge node;

[0018] Based on the first relevance of each knowledge node and the predefined prior attention trend, the interest mapping factor of the corresponding knowledge node is determined.

[0019] In one possible implementation of the first aspect, determining the knowledge enhancement coefficient of the corresponding knowledge node based on the first relevance of each knowledge node includes:

[0020] Obtain predefined knowledge reinforcement functions;

[0021] The knowledge clustering degree of knowledge reinforcement processing in this round of iterative optimization is determined; the knowledge clustering degree is a feature value determined based on the prior optimized health and wellness interest graph, which is the optimized health and wellness interest graph generated in the previous round of iterative optimization before this round of iterative optimization; the knowledge clustering degree includes a first knowledge clustering degree and a second knowledge clustering degree, wherein the first knowledge clustering degree is greater than the second knowledge clustering degree; when iteratively optimizing the first health and wellness interest graph, the knowledge reinforcement coefficient is iteratively determined based on the first knowledge clustering degree and the second knowledge clustering degree.

[0022] The knowledge aggregation degree and the first correlation degree of each knowledge node are loaded into the knowledge reinforcement function to generate the knowledge reinforcement coefficient of the corresponding knowledge node.

[0023] In one possible implementation of the first aspect, determining the interest mapping factor of the corresponding knowledge node based on the first relevance of each knowledge node and a predefined prior attention trend includes:

[0024] Obtain the health preservation association vector that maps each knowledge node to the TCM health preservation knowledge profile node of the target user, and determine the deviation value between the health preservation association vector and the prior attention trend;

[0025] Obtain a predefined interest mapping function and determine candidate adjustment parameters for the interest mapping in the current iteration of optimization. The candidate adjustment parameters are feature values ​​determined based on the prior optimized health and wellness interest map, which is the optimized health and wellness interest map generated in the previous iteration of optimization before the current iteration of optimization.

[0026] The candidate adjustment parameters and the deviation values ​​are loaded into the interest mapping function to generate the interest mapping factors for the corresponding knowledge nodes.

[0027] In one possible implementation of the first aspect, the step of performing knowledge enhancement processing on all knowledge nodes in the TCM health preservation knowledge domain based on the knowledge enhancement coefficient to generate multiple extended mining nodes includes:

[0028] Based on the knowledge enhancement coefficient, the knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing in the prior attention trend corresponding to the tracking parameters, thereby generating the multiple extended mining nodes.

[0029] In one possible implementation of the first aspect, the step of performing knowledge enhancement processing on knowledge nodes in the TCM health preservation knowledge domain based on the knowledge enhancement coefficient and the prior attention trend corresponding to the tracking parameters to generate the plurality of extended mining nodes includes:

[0030] For each of the multiple prior attention trends corresponding to the tracking parameters, obtain the knowledge reinforcement coefficient of each of the multiple knowledge nodes.

[0031] Knowledge nodes whose knowledge enhancement coefficient is greater than the threshold value are output as extended mining nodes on the prior attention trend.

[0032] In one possible implementation of the first aspect, the step of performing interest mapping on the plurality of extended mining nodes based on the interest mapping factor to generate an interest mapping vector for each of the extended mining nodes includes:

[0033] The multiple health feedback data are encoded and represented to generate a first encoded representation vector for each of the extended mining nodes;

[0034] Obtain the first encoded representation vector of multiple extended mining nodes on each prior interest trend and the interest mapping factor of each extended mining node;

[0035] Determine the fusion calculation result between the interest mapping factor of each of the extended mining nodes and the first encoded representation vector;

[0036] Based on the prior attention trend, the fusion calculation results are mapped to interests, and the fusion calculation results of each extended mining node are accumulated to generate an interest mapping vector that maps to the TCM health preservation knowledge profile of the target user.

[0037] In one possible implementation of the first aspect, the step of iteratively optimizing the first health-preserving interest graph of the target user based on the interest mapping vector of the extended mining node to generate an optimized health-preserving interest graph in the field of traditional Chinese medicine health-preserving knowledge of the target user includes:

[0038] Obtain the interest mapping vector and actual interest vector of the extended mining node in the target user's TCM health preservation knowledge profile;

[0039] Determine the feature distance between the interest mapping vector and the actual interest vector;

[0040] The error parameters of the TCM health preservation knowledge association model are determined based on the feature distance; the TCM health preservation knowledge association model is used to predict each first association degree in the first health preservation interest graph.

[0041] Based on the error parameters, the model parameter information of the TCM health preservation knowledge association model is optimized to generate an optimized TCM health preservation knowledge association model;

[0042] The optimized TCM health preservation knowledge association model is used to iteratively optimize the first health preservation interest graph of the target user, thereby generating an optimized health preservation interest graph in the TCM health preservation knowledge domain of the target user.

[0043] The step of constructing a TCM health preservation plan for the target user based on the optimized health preservation interest graph and generating a TCM health preservation plan recommended to the target user includes:

[0044] The optimized health and wellness interest graph is analyzed to identify various knowledge nodes and the relationships between them, generating a basic framework for a health and wellness plan that includes multiple health and wellness modules.

[0045] For each health preservation section, the depth and breadth of health preservation information contained in the corresponding knowledge nodes in the optimized health preservation interest graph are analyzed to obtain knowledge node information. Based on the knowledge node information, a preliminary plan is made for each health preservation section, generating the preliminary plan content for each health preservation section. The preliminary plan content of each health preservation section is integrated to form a complete TCM health preservation plan.

[0046] In another aspect, embodiments of the present invention also provide a personalized recommendation system for TCM health preservation plans based on big data, including a processor and a machine-readable storage medium. The machine-readable storage medium is connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code. The processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the above-mentioned method.

[0047] Based on the above, this application embodiment constructs a first health preservation interest graph for the target user, accurately capturing the user's interests and needs in the field of traditional Chinese medicine (TCM) health preservation knowledge. Utilizing knowledge reinforcement coefficients and interest mapping factors, it deeply mines and personalizes knowledge nodes in the TCM health preservation knowledge field, generating an optimized health preservation interest graph. This not only improves the accuracy and personalization of TCM health preservation plan recommendations but also continuously iterates and optimizes the recommendation results through a cyclical optimization mechanism, effectively meeting the TCM health preservation needs of the target user and improving the practicality and user satisfaction of the health preservation plans. Attached Figure Description

[0048] Figure 1 This is a schematic diagram of the execution flow of the personalized recommendation method for TCM health preservation based on big data provided in the embodiments of the present invention.

[0049] Figure 2 This is a schematic diagram of the hardware architecture of the personalized recommendation system for TCM health preservation based on big data provided in this embodiment of the invention. Detailed Implementation

[0050] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a method for personalized recommendation of TCM health preservation plans based on big data, provided in one embodiment of the present invention. The following is a detailed description of this method.

[0051] Step S110: Based on multiple health feedback data of the target user and the tracking parameters corresponding to each health feedback data, construct a first health interest graph in the TCM health knowledge domain of the target user. The first health interest graph includes a first correlation degree between each knowledge node in the TCM health knowledge domain and the TCM health knowledge profile of the target user.

[0052] In this embodiment, it is assumed that the server receives health feedback data from the target user. This health feedback data comes from the user's daily health monitoring devices, such as sleep data and exercise data recorded by a smart bracelet, as well as dietary records and physical discomfort symptoms that the user regularly fills in on a health platform. Each health feedback data has corresponding tracking parameters. For example, the tracking parameters for sleep data may include sleep time, sleep quality score, and deep sleep duration; the tracking parameters for exercise data may include exercise type, exercise duration, and exercise intensity; and the tracking parameters for dietary records may include food type and intake.

[0053] The server first extracts information on the scope of health preservation knowledge and monitoring parameters from these tracking parameters within the field of Traditional Chinese Medicine (TCM) health preservation. For example, by analyzing a user's dietary records, if it is found that the user consumes a lot of spicy food and is prone to "internal heat," the scope of health preservation information may involve the TCM concept of "internal heat," falling under the category of dietary health preservation within the TCM health preservation knowledge field. The health preservation monitoring parameters might include the specific amount and frequency of spicy food intake.

[0054] Based on the scope of health preservation information, the target user's TCM health preservation knowledge domain is determined. For example, if a diet is related to "internal heat" as mentioned above, it falls under the TCM dietary health preservation domain. Then, based on health preservation monitoring parameters, the attribute information of each knowledge node within the TCM health preservation knowledge domain is determined. In the TCM dietary health preservation domain, knowledge nodes may include the nature and flavor of food (e.g., cold, hot), and the effects of food (e.g., clearing heat and reducing fire, nourishing yin and moistening dryness). The attribute information is related to the user's specific dietary data; for example, chili peppers, which the user frequently eats, are considered hot in nature and may cause internal heat.

[0055] The server loads this attribute information and multiple health feedback data into a pre-built TCM health preservation knowledge association model. This model contains a large amount of TCM health preservation knowledge and the relationships between them. For example, the model stores the association between cold-natured foods and clearing heat and reducing internal fire, as well as the adaptability relationship with different body constitutions. By loading the user's chili pepper intake (attribute information) and other health feedback data (such as symptoms of being prone to internal heat) into the model, the first degree of association between each knowledge node in the TCM health preservation knowledge domain and the target user's TCM health preservation knowledge profile can be generated. For example, the chili pepper knowledge node may have a high degree of association with the user's TCM health preservation knowledge profile because it is closely related to the user's symptoms of internal heat; while the cold-natured food knowledge node, such as mung beans, may have a low degree of association with the user profile because the user does not show a related need. Finally, these associations are used to construct the first health preservation interest graph in the TCM health preservation knowledge domain of the target user. This graph is like a map, showing the degree of association between each knowledge node (such as different foods, health preservation methods, etc.) and the user's TCM health preservation knowledge profile.

[0056] Step S120: Based on the first relevance of each knowledge node and the predefined prior attention trend, determine the knowledge reinforcement coefficient and interest mapping factor of the corresponding knowledge node respectively.

[0057] In this embodiment, this step begins after the server has constructed the first health and wellness interest graph. Let's first look at the part about determining the knowledge reinforcement coefficient. Assume the predefined knowledge reinforcement function is (f(x, y)), where (x) represents the knowledge clustering degree and (y) represents the first correlation degree. In this round of iterative optimization, the server needs to determine the knowledge clustering degree for knowledge reinforcement processing. This knowledge clustering degree is a feature value determined based on the prior optimized health and wellness interest graph, which is the optimized health and wellness interest graph generated in the previous round of iterative optimization. For example, suppose the prior optimized health and wellness interest graph shows that in the field of dietary health and wellness, the clustering degree of knowledge nodes related to food nutritional components is high, while the clustering degree of knowledge nodes related to food cooking methods is low. The knowledge clustering degree here includes the first knowledge clustering degree and the second knowledge clustering degree. The first knowledge clustering degree is greater than the second knowledge clustering degree. For example, the first knowledge clustering degree might be the clustering degree for core health and wellness knowledge nodes, and the second knowledge clustering degree might be the clustering degree for auxiliary health and wellness knowledge nodes. When iteratively optimizing the first health-related interest graph, the server determines the knowledge enhancement coefficient based on the first and second knowledge clustering degrees. For example, for a chili pepper knowledge node that is highly correlated with a user's symptoms of internal heat (first correlation degree), if it is located in a region with previously high knowledge clustering degree (the region corresponding to the first knowledge clustering degree), then loading the knowledge clustering degree and the first correlation degree of this knowledge node into the knowledge enhancement function (f(x, y)) may generate a higher knowledge enhancement coefficient; while for a mung bean knowledge node with low correlation degree and located in a region with low knowledge clustering degree, the generated knowledge enhancement coefficient will be lower.

[0058] Next, let's look at the part that determines the interest mapping factors. The server obtains the health-related vector of each knowledge node mapped to the target user's TCM health-preserving knowledge profile node. For example, for the chili pepper knowledge node, its health-preserving related vector may include the correlation between its stimulating effect on the body, its impact on the digestive system, and the user's health-preserving knowledge profile nodes (such as constitution prone to internal heat, gastrointestinal function, etc.). The deviation value between this health-preserving related vector and the prior attention trend is determined. Suppose the prior attention trend is to reduce the intake of stimulating foods to improve symptoms of internal heat, and the health-preserving related vector of chili pepper indicates that it is a stimulating food, then the deviation value will be relatively large. Simultaneously, the server obtains the predefined interest mapping function (g(x, y)) and determines the candidate adjustment parameters for the interest mapping in this round of iterative optimization. These candidate adjustment parameters are also feature values ​​determined based on the prior optimized health-preserving interest graph. For example, if the prior optimized health-preserving interest graph shows a high sensitivity to food adjustments in improving symptoms of internal heat, this sensitivity is part of the candidate adjustment parameters. The candidate adjustment parameters and deviation values ​​are then loaded into the interest mapping function (g(x, y)) to generate the interest mapping factor for the corresponding knowledge node. For the chili pepper knowledge node, due to the large deviation value and the influence of the candidate adjustment parameters, a large interest mapping factor may be generated, indicating that it will receive more attention and adjustment in subsequent processing.

[0059] Step S130: Based on the knowledge enhancement coefficient, perform knowledge enhancement processing on all knowledge nodes in the TCM health preservation knowledge domain to generate multiple extended mining nodes, and perform interest mapping on the multiple extended mining nodes based on the interest mapping factor to generate an interest mapping vector for each extended mining node.

[0060] In this embodiment, the server performs knowledge enhancement processing on knowledge nodes in the TCM health preservation knowledge domain based on the knowledge enhancement coefficient within the prior interest trends corresponding to the tracking parameters, generating multiple extended mining nodes. For example, for multiple knowledge nodes on each of the multiple prior interest trends corresponding to the tracking parameters, the server first examines the knowledge nodes related to the prior interest trend of dietary health preservation. The server obtains the knowledge enhancement coefficient of each of these knowledge nodes. For the previously mentioned chili pepper knowledge node, its knowledge enhancement coefficient is relatively high. Assuming the threshold value is set to 0.5, the knowledge enhancement coefficient of the chili pepper knowledge node is 0.8, which is greater than the threshold value, so the server outputs it as an extended mining node on the prior interest trend (dietary health preservation). However, the knowledge enhancement coefficient of the mung bean knowledge node is 0.3, which is less than the threshold value, so it is not output as an extended mining node.

[0061] Next, the server performs interest mapping on multiple extended mining nodes based on the interest mapping factor, generating an interest mapping vector for each extended mining node. The server encodes multiple health feedback data, generating a first encoded representation vector for each extended mining node. For example, for the chili pepper extended mining node, user health feedback data regarding symptoms of internal heat (such as the degree of oral inflammation, constipation, etc.) is encoded to form a first encoded representation vector. Simultaneously, the server obtains the first encoded representation vectors and interest mapping factors for multiple extended mining nodes under each prior interest trend. For instance, under the prior interest trend of dietary health, besides chili peppers, there are other extended mining nodes such as ginger (assuming ginger also meets the criteria to become an extended mining node), and their respective first encoded representation vectors and interest mapping factors are obtained. Then, the server determines the fusion calculation result between the interest mapping factor and the first encoded representation vector for each extended mining node. For the chili pepper extended mining node, its interest mapping factor may undergo a special fusion calculation with the encoding part related to symptoms of internal heat in the first encoded representation vector, which may be a weighted sum or other specific calculation method. Finally, based on prior interest trends, interest mapping is performed on the fusion calculation results. The fusion calculation results of each extended mining node are accumulated separately to generate an interest mapping vector that maps to the target user's TCM health knowledge profile. For example, under the prior interest trend of dietary health, the fusion calculation results of extended mining nodes such as chili peppers and ginger are accumulated according to certain rules to finally obtain an interest mapping vector that maps to the user's TCM health knowledge profile. This interest mapping vector can reflect the comprehensive influence of these extended mining nodes on the user's health.

[0062] Step S140: Based on the interest mapping vector of the extended mining node, perform iterative optimization on the first health preservation interest map of the target user to generate an optimized health preservation interest map in the field of traditional Chinese medicine health preservation knowledge of the target user.

[0063] In this embodiment, the server first obtains the interest mapping vector and actual interest vector of the extended mining node in the target user's TCM health knowledge profile. For example, for the previous chili pepper extended mining node, the interest mapping vector may reflect its theoretical influence on the user's health, while the actual interest vector may be obtained by analyzing the user's actual dietary choices (even though chili peppers may have an effect on causing internal heat, the user still eats them frequently). Then, the feature distance between the interest mapping vector and the actual interest vector is determined. If the interest mapping vector indicates that chili peppers should be eaten as little as possible to improve symptoms of internal heat, but the actual interest vector shows that the user eats them frequently, then this feature distance is relatively large.

[0064] The error parameters of the TCM health preservation knowledge association model are determined based on this feature distance. This model is used to predict each first association degree in the first health preservation interest graph. For example, a large feature distance indicates a significant deviation between the model's predicted association degree and the actual situation, resulting in a larger error parameter. Next, the model parameters of the TCM health preservation knowledge association model are optimized based on the error parameter, generating an optimized model. For example, the weight parameters related to the relationship between food and constitution may be adjusted.

[0065] Finally, the optimized TCM health preservation knowledge association model is used to iteratively optimize the target user's primary health preservation interest graph, generating an optimized health preservation interest graph within the target user's TCM health preservation knowledge domain. In this process, the optimized health preservation interest graph is analyzed to identify various knowledge nodes and the relationships between them, generating a basic framework for health preservation solutions containing multiple modules. For example, dietary health preservation, exercise health preservation, and emotional health preservation modules are identified, and the relationships between knowledge nodes (such as food, exercise methods, and emotion regulation methods) within each module are clarified. For each health preservation module, the depth and breadth of health preservation information contained in the corresponding knowledge nodes in the optimized health preservation interest graph are analyzed to obtain knowledge node information. For example, in the dietary health preservation module, the depth and breadth of information such as nutritional components, efficacy, and applicable populations for different food knowledge nodes are analyzed. Based on the knowledge node information, preliminary planning is carried out for each health preservation module, generating preliminary planning content for each module. For example, the initial plan in the dietary health section includes increasing the intake of cold foods and reducing the intake of hot foods. The initial plans of various health sections will be integrated to form a complete TCM health program.

[0066] Step S150: Based on the optimized health preservation interest map, construct a TCM health preservation plan for the target user and generate a TCM health preservation plan recommended to the target user.

[0067] In this embodiment, the server constructs a TCM health preservation plan based on the previously generated optimized health preservation interest graph. Building upon the basic framework of the health preservation plan, the content of each health preservation section is further refined. Taking the dietary health preservation section as an example, specific food recommendations are determined based on the correlation between food knowledge nodes in the optimized health preservation interest graph and the user's health preservation knowledge profile, as well as other relevant information. If a user is prone to internal heat and has a constitution that leans towards heat, and the correlation between cold foods and the user's profile is high in the optimized health preservation interest graph, then the dietary health preservation section of the health preservation plan will provide detailed recommendations on the intake and frequency of consumption of cold foods such as bitter melon and winter melon.

[0068] For the exercise and wellness section, based on the information from the exercise knowledge nodes in the optimized wellness interest graph—such as the correlation between the effects of different exercise methods (e.g., Tai Chi, jogging) on ​​the body's blood circulation and physical conditioning and the user's wellness needs—the appropriate exercise methods, durations, and intensities are determined. If the graph shows a high correlation between Tai Chi and the user's ability to regulate blood circulation and soothe emotions, then the wellness plan will recommend that the user practice Tai Chi a certain number of times and for a certain duration each week.

[0069] In the section on emotional well-being, based on the knowledge nodes related to the relationship between emotions and the body's internal organs in the optimized wellness interest map, such as the correlation between long-term anxiety and liver function, suggestions for emotional regulation are provided to users. If a user is prone to anxiety due to high work stress, the wellness plan may suggest that the user regulate their emotions through meditation, listening to soothing music, etc., in order to protect liver health.

[0070] By meticulously planning each health and wellness section, the server ultimately generates a complete TCM health and wellness plan tailored to the target user. This plan comprehensively considers the user's health feedback data, various knowledge nodes in the field of TCM health and wellness, and the relationships between them, providing users with personalized TCM health and wellness guidance.

[0071] Based on the above steps, this embodiment of the application constructs a first health preservation interest graph for the target user, accurately capturing the user's interests and needs in the field of traditional Chinese medicine (TCM) health preservation knowledge. Utilizing knowledge reinforcement coefficients and interest mapping factors, it deeply mines and personalizes knowledge nodes in the TCM health preservation knowledge field, generating an optimized health preservation interest graph. This not only improves the accuracy and personalization of TCM health preservation plan recommendations but also continuously iterates and optimizes the recommendation results through a cyclical optimization mechanism, effectively meeting the TCM health preservation needs of the target user and improving the practicality and user satisfaction of the health preservation plans.

[0072] In one possible implementation, step S110 includes:

[0073] Step S111: Determine the TCM health knowledge domain of the target user based on the health scope information.

[0074] Step S112: Based on the health monitoring parameters, determine the attribute information of each knowledge node in the TCM health knowledge domain.

[0075] Step S113: Load the attribute information and the multiple health feedback data into the TCM health preservation knowledge association model to generate the first association degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user.

[0076] Step S114: Using the first correlation degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user, construct the first health preservation interest map of the TCM health preservation knowledge domain in which the target user is located.

[0077] In this embodiment, firstly, information on the scope of health preservation knowledge and health monitoring parameters in the field of Traditional Chinese Medicine (TCM) health preservation are extracted from the tracking parameters. For example, the target user provides rich health feedback data through smart devices and health record platforms. This data includes information such as daily steps, heart rate changes, types and amounts of food consumed, sleep duration and quality, and each data point has corresponding tracking parameters. From the step tracking parameters, the server can extract health preservation information related to exercise intensity. For example, if the number of steps is consistently high and the pace is fast, it may fall under the category of moderate physical exertion in TCM health preservation. The corresponding health monitoring parameters for steps may be the specific step count and the step distribution over different time periods. From the perspective of diet-related tracking parameters, if the user records the types and amounts of vegetables, meat, grains, etc., consumed daily, the health preservation information may involve nutritional balance in TCM dietary health preservation. The health monitoring parameters are the actual intake and frequency of various foods.

[0078] Next, the target user's TCM (Traditional Chinese Medicine) health knowledge domain is determined based on the health scope information. As mentioned earlier, if the health scope information contains a lot of content related to exercise intensity, number of steps, etc., and points to the effects of moderate exercise on the body, then the user is likely in the TCM exercise health knowledge domain. If the health scope information is dominated by diet-related information, such as the types of food and nutritional combinations, then it is determined to be in the TCM diet health knowledge domain. For example, if the user's diet record shows that they consume a lot of greasy food and have low dietary fiber intake, and the health scope information revolves around an unreasonable diet, then the user is clearly in the TCM diet health knowledge domain.

[0079] Then, based on the health monitoring parameters, the attribute information of each knowledge node in the field of Traditional Chinese Medicine (TCM) health preservation is determined. In the field of TCM dietary health preservation, if the health monitoring parameters show that the user consumes a large amount of greasy food, the attribute information for the knowledge node of greasy food might be high in fat, high in calories, and difficult to digest. This attribute information is based on TCM's understanding of food characteristics. For example, greasy food may increase the burden on the spleen and stomach, which is related to knowledge related to spleen and stomach function in TCM theory. Similarly, if the user has short sleep duration and poor sleep quality, in the field of TCM sleep health preservation, the attribute information for the knowledge node of sleep would be insufficient sleep time and potential impact on the regulation of Qi and blood. This is because sleep and Qi and blood are closely related in TCM, and insufficient sleep may affect the abundance and circulation of Qi and blood.

[0080] Next, attribute information and multiple health feedback data are loaded into the TCM health preservation knowledge association model to generate the first degree of association between each knowledge node in the TCM health preservation knowledge domain and the target user's TCM health preservation knowledge profile. This TCM health preservation knowledge association model is a knowledge base containing a large amount of TCM health preservation knowledge and their interrelationships. For example, for the knowledge node of greasy food, its attribute information (high in calories, difficult to digest, etc.) and the user's health feedback data (such as being overweight, occasional indigestion symptoms, etc.) are loaded into the model. The model stores information such as the relationship between greasy food and weight, spleen and stomach function in TCM knowledge. By analyzing this information, the degree of association between the greasy food knowledge node and the target user's TCM health preservation knowledge profile can be obtained. If the user has indigestion symptoms and is overweight, and greasy food easily causes these problems, then the degree of association between them will be relatively high. For knowledge nodes of healthy foods such as vegetables, because the user consumes less and has health problems, the degree of association between them and the user's TCM health preservation knowledge profile may be relatively low, because the user's current health condition is related to a lack of vegetable intake, but this is not taken seriously in actual behavior.

[0081] Finally, by leveraging the primary correlation between each knowledge node in the TCM health preservation knowledge domain and the target user's TCM health preservation knowledge profile, a primary health preservation interest graph is constructed within the target user's TCM health preservation knowledge domain. This graph acts like a map showcasing the relationship between the user and TCM health preservation knowledge. Taking the TCM dietary health preservation domain as an example, the knowledge node on oily foods in the graph is likely to be prominent due to its high correlation with the user's TCM health preservation knowledge profile, surrounded by related health issues (such as being overweight or having indigestion) and related concepts in TCM theory (such as spleen and stomach function). Conversely, the knowledge node on vegetables, due to its low correlation, is relatively peripheral in the graph. This primary health preservation interest graph provides a foundation for subsequent TCM health preservation analysis and plan development for the user. Through this graph, the close relationship between each knowledge node and the user's health status can be visually observed, thus providing a basis for further adjustments to health preservation strategies.

[0082] In one possible implementation, step S120 includes:

[0083] Step S121: Determine the knowledge reinforcement coefficient of the corresponding knowledge node based on the first correlation degree of each knowledge node.

[0084] Step S122: Determine the interest mapping factor of the corresponding knowledge node based on the first relevance of each knowledge node and the predefined prior attention trend.

[0085] In one possible implementation, step S121 includes:

[0086] Step S1211: Obtain the predefined knowledge reinforcement function.

[0087] Step S1212: Determine the knowledge clustering degree of the knowledge reinforcement processing in this round of iterative optimization. The knowledge clustering degree is a feature value determined based on the prior optimized health and wellness interest graph, which is the optimized health and wellness interest graph generated in the previous round of iterative optimization. The knowledge clustering degree includes a first knowledge clustering degree and a second knowledge clustering degree, where the first knowledge clustering degree is greater than the second knowledge clustering degree. During iterative optimization of the first health and wellness interest graph, the knowledge reinforcement coefficient is iteratively determined based on the first knowledge clustering degree and the second knowledge clustering degree.

[0088] Step S1213: Load the knowledge aggregation degree and the first correlation degree of each knowledge node into the knowledge reinforcement function to generate the knowledge reinforcement coefficient of the corresponding knowledge node.

[0089] In this embodiment, the server first determines the knowledge enhancement coefficient of each knowledge node based on its first relevance. During this process, the server first obtains a predefined knowledge enhancement function. This function is pre-defined and is a mathematical expression constructed based on the TCM health preservation knowledge system and the analysis results of a large amount of past user data. It aims to calculate suitable knowledge enhancement coefficients based on different input parameters. For example, this function may be a complex multivariate function that comprehensively considers multiple factors affecting knowledge enhancement.

[0090] Next, the server needs to determine the knowledge clustering degree of the knowledge reinforcement processing in this round of iterative optimization. This knowledge clustering degree is a feature value determined based on the prior optimized health care interest graph, which is the optimized health care interest graph generated in the previous round of iterative optimization. Taking the aforementioned field of traditional Chinese medicine dietary health care as an example, suppose the prior optimized health care interest graph shows that the clustering degree of knowledge nodes related to the nutritional components of food is high, while the clustering degree of knowledge nodes related to the cooking methods of food is low. The knowledge clustering degree here includes the first knowledge clustering degree and the second knowledge clustering degree, with the first knowledge clustering degree being greater than the second knowledge clustering degree. The first knowledge clustering degree may represent the degree of clustering of more core and critical knowledge nodes in the health care system. For example, in traditional Chinese medicine dietary health care, the clustering degree of knowledge nodes related to the basic nutritional characteristics of food, such as its nature, flavor, and meridian tropism, may be classified as the first knowledge clustering degree because these characteristics directly affect the effect of food on human health. The second knowledge clustering degree may be the degree of clustering of relatively minor but still influential knowledge nodes, such as the clustering degree of knowledge nodes related to information such as the color and origin of food. When iterating and optimizing the first health interest graph, the server determines the knowledge reinforcement coefficient based on the first knowledge clustering degree and the second knowledge clustering degree.

[0091] Taking a specific knowledge node as an example, suppose there is a knowledge node called "greasy food" in the field of Traditional Chinese Medicine (TCM) diet and health preservation. If the knowledge node "greasy food" has a high degree of first-order relevance to the target user's TCM health preservation knowledge profile in the previous first health preservation interest graph, and it is located in a region with high first-order knowledge clustering in the prior optimized health preservation interest graph, for example, it is closely related to the food nutrition knowledge node (because the high fat and other nutrients in greasy food are important characteristics), the server loads the knowledge clustering (here, a high first-order knowledge clustering) and the first-order relevance of the greasy food knowledge node into the knowledge reinforcement function. Assuming the knowledge reinforcement function is (f(x, y)), where (x) represents the knowledge clustering and (y) represents the first-order relevance, the corresponding knowledge reinforcement coefficient is generated by calculating (f(high knowledge clustering, high first-order relevance)). This coefficient reflects the degree to which the greasy food knowledge node should be reinforced in the subsequent health preservation knowledge optimization process. Because of its high relevance and location in a high knowledge clustering region, the generated knowledge reinforcement coefficient is likely to be relatively high.

[0092] In one possible implementation, step S122 includes:

[0093] Step S1221: Obtain the health preservation association vector of each knowledge node mapped to the TCM health preservation knowledge profile node of the target user, and determine the deviation value between the health preservation association vector and the prior attention trend.

[0094] Step S1222: Obtain the predefined interest mapping function and determine the candidate adjustment parameters for the interest mapping in this round of iterative optimization. The candidate adjustment parameters are feature values ​​determined based on the prior optimized health care interest map, which is the optimized health care interest map generated in the previous round of iterative optimization before this round of iterative optimization.

[0095] Step S1223: Load the candidate adjustment parameters and the deviation value into the interest mapping function to generate the interest mapping factor of the corresponding knowledge node.

[0096] On the other hand, the server determines the interest mapping factor for each knowledge node based on its initial relevance and predefined prior attention trends. The server first obtains the health-related vector of each knowledge node mapped to the target user's TCM health-preserving knowledge profile node. Taking the knowledge node of oily food in the field of TCM dietary health preservation as an example, its health-related vector may include multiple aspects such as its impact on blood lipid levels, its burden on the spleen and stomach's digestive function, and its association with obesity. These aspects are related to the target user's TCM health-preserving knowledge profile nodes (such as the user's blood lipid status, spleen and stomach function, weight, etc.). Then, the deviation value between this health-related vector and the prior attention trend is determined. If the prior attention trend is towards improving the user's blood lipid levels, reducing the burden on the spleen and stomach, and controlling weight, while the health-related vector of oily food indicates that it has an effect on increasing blood lipids, increasing the burden on the spleen and stomach, and may lead to weight gain, then this deviation value will be relatively large.

[0097] Simultaneously, the server needs to obtain a predefined interest mapping function and determine candidate adjustment parameters for the interest mapping in the current iteration of optimization. This interest mapping function is also pre-constructed and used to calculate interest mapping factors based on the input parameters. The candidate adjustment parameters are feature values ​​determined based on a prior optimized health and wellness interest graph, which is an optimized health and wellness interest graph generated in the previous iteration before the current iteration. For example, the prior optimized health and wellness interest graph might show a high sensitivity to food adjustments in improving blood lipids; this sensitivity could become part of the candidate adjustment parameters. Assuming the interest mapping function is (g(x, y)), where (x) represents the candidate adjustment parameter and (y) represents the deviation value, the server loads the candidate adjustment parameters (such as food adjustment sensitivity) and the deviation value (the deviation between the oily food health association vector and the prior attention trend) into the interest mapping function (g(x, y)), thus generating the interest mapping factors for the corresponding knowledge nodes (such as the oily food knowledge node). Because of the significant deviation between oily food and prior interest trends, and the influence of candidate adjustment parameters (such as high sensitivity to food adjustment), the generated interest mapping factor may be relatively large. This indicates that in the subsequent processing of health knowledge, the knowledge node of oily food needs to be adjusted more according to the interest mapping factor in order to better meet health goals.

[0098] Taking the knowledge node of exercise intensity in the field of traditional Chinese medicine exercise and health preservation as an example, if the knowledge node of exercise intensity has a low first correlation with the target user's TCM health preservation knowledge profile in the first health preservation interest graph, and it is located in a region with low second knowledge clustering in the prior optimized health preservation interest graph, for example, it is closely related to the relatively minor environmental factor knowledge node in exercise and health preservation, then loading this low knowledge clustering and low first correlation into the knowledge reinforcement function (f(x, y)) may result in a relatively low knowledge reinforcement coefficient.

[0099] The determination of the interest mapping factor for the exercise intensity knowledge node involves considering its health-related vector, which may include the relationship between the impact on physical fatigue recovery and cardiovascular system stress, and the target user's TCM health-related knowledge profile nodes (such as the user's fatigue recovery ability and cardiovascular health status). Assuming the prior interest trend is towards improving fatigue recovery ability and reducing cardiovascular system stress, if the exercise intensity is too high, the health-related vector will deviate from the prior interest trend. The server obtains the interest mapping function (g(x, y)) and determines candidate adjustment parameters for the interest mapping based on the prior optimized health-related interest graph (such as the sensitivity of exercise intensity adjustment in improving cardiovascular health). The candidate adjustment parameters and deviation values ​​are loaded into the interest mapping function to generate the interest mapping factor for the exercise intensity knowledge node. If the deviation value is small and the sensitivity in the candidate adjustment parameters is low, the generated interest mapping factor may be relatively small, meaning that the adjustment range of this knowledge node will be relatively small in subsequent health-related knowledge processing.

[0100] Throughout the entire TCM health preservation knowledge system, the process of determining the knowledge enhancement coefficient and interest mapping factor for different knowledge nodes is similar. In this way, the server can accurately determine the knowledge enhancement coefficient and interest mapping factor for each knowledge node based on factors such as the relevance of the knowledge node to the user's health profile and prior attention trends. This provides an important basis for subsequent optimization of health preservation knowledge and the development of personalized health preservation plans. For example, in the field of TCM emotional health preservation, the knowledge node on emotional regulation methods (such as regulating anxiety through meditation) is determined based on factors such as its primary relevance to the user's health profile (e.g., the severity of the user's anxiety), the knowledge aggregation degree in the prior optimized health preservation interest graph (e.g., the importance of emotional regulation in the entire health preservation system), and prior attention trends (e.g., moving towards relieving anxiety and improving mental health). In this way, the server can then use these knowledge enhancement coefficients and interest mapping factors to target knowledge nodes in different health preservation knowledge areas to better meet the health preservation needs of the target users.

[0101] In one possible implementation, step S130 includes:

[0102] Based on the knowledge enhancement coefficient, the knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing in the prior attention trend corresponding to the tracking parameters, thereby generating the multiple extended mining nodes.

[0103] In one possible implementation, based on the knowledge enhancement coefficient, knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing in the prior attention trend corresponding to the tracking parameters to generate the multiple extended mining nodes, including:

[0104] Step S131: For each of the multiple prior attention trends corresponding to the tracking parameters, obtain the knowledge reinforcement coefficient of each of the multiple knowledge nodes.

[0105] Step S132: Output the knowledge nodes whose knowledge enhancement coefficient is greater than the threshold value as the extended mining nodes on the prior attention trend.

[0106] In this embodiment, firstly, the server, based on the knowledge enhancement coefficient, performs knowledge enhancement processing on the knowledge nodes in the TCM health preservation knowledge domain within the prior attention trends corresponding to the tracking parameters to generate multiple extended mining nodes. The prior attention trends corresponding to the tracking parameters are directions determined based on previous analysis of user health data and health preservation goals. Taking the TCM dietary health preservation domain as an example, tracking parameters may include food types, intake amounts, and nutritional components, while prior attention trends may develop towards improving the user's blood lipid levels, regulating spleen and stomach function, or controlling weight.

[0107] Then, for each of the multiple prior interest trends corresponding to the tracking parameters, the server obtains the knowledge reinforcement coefficient for each of these knowledge nodes. For example, in the field of traditional Chinese medicine dietary health preservation, under the prior interest trend of improving blood lipid levels, knowledge nodes may include various foods such as greasy foods (e.g., fried foods), high-fiber foods (e.g., oats), foods rich in unsaturated fatty acids (e.g., fish oil), and may also include knowledge nodes such as regular eating habits (e.g., eating small meals frequently). For each such knowledge node, the server has already determined its knowledge reinforcement coefficient based on previous calculations. For the greasy food knowledge node, since it may have adverse effects on blood lipids and is closely related to the user's health preservation in previous analysis (e.g., the user has high blood lipids and frequently consumes greasy foods), its knowledge reinforcement coefficient may be relatively high; while for the high-fiber food knowledge node, since it helps improve blood lipids and may be lacking in the user's diet (assuming the user's dietary fiber intake is insufficient), its knowledge reinforcement coefficient may also be high; for the regular eating habits knowledge node, if the user's diet is irregular and this has a certain impact on blood lipids, its knowledge reinforcement coefficient may also be at a high level.

[0108] Finally, the server outputs knowledge nodes with a knowledge enhancement coefficient greater than the threshold as expanded mining nodes based on the prior interest trend. Assuming the threshold is set at 0.6, for the knowledge node about oily foods, if its knowledge enhancement coefficient is 0.8, which is greater than the threshold, the server will output this knowledge node as an expanded mining node based on the prior interest trend of improving blood lipid levels. This means that in subsequent health knowledge optimization and health plan formulation, the knowledge node about oily foods will be given priority and its related knowledge will be further explored. For example, further analysis of the specific mechanisms by which oily foods affect blood lipids, how to reduce the intake of oily foods, or how to choose relatively healthy alternatives to oily foods will be conducted. For the knowledge node about high-fiber foods, if its knowledge enhancement coefficient is 0.7, which is also greater than the threshold, it will also be output as an expanded mining node. At this point, for high-fiber foods, further exploration may be conducted on the differences in the effects of different types (such as oats, whole wheat, etc.) on improving blood lipids, the optimal intake, and how to combine them with other foods to better achieve the effect of improving blood lipids. For knowledge nodes related to dietary patterns, if their knowledge reinforcement coefficient is 0.65, which is greater than the threshold, they will also become nodes for further exploration, allowing for in-depth research into specific patterns of eating smaller, more frequent meals suitable for users, such as the interval between meals and the approximate proportion of food intake per meal.

[0109] Taking the field of Traditional Chinese Medicine (TCM) exercise and wellness as an example, the tracking parameters might include exercise type, intensity, and duration. The prior focus trend hypothesis is improving cardiopulmonary function and enhancing flexibility. Knowledge nodes might include aerobic exercise (such as jogging and swimming), strength training (such as weightlifting and push-ups), and flexibility training (such as yoga and Pilates). For the aerobic exercise knowledge node, if the user has weak cardiopulmonary function and previous analysis indicates that aerobic exercise has significant potential to improve it, its knowledge reinforcement coefficient may be high. For the strength training knowledge node, if its effect on improving cardiopulmonary function is relatively small and its relevance to the user's current needs is low, its knowledge reinforcement coefficient may be low. For the flexibility training knowledge node, if it directly helps improve flexibility and meets the user's wellness needs (such as the user having poor flexibility and wanting to improve it), its knowledge reinforcement coefficient may be high. Assuming a threshold of 0.5, and the knowledge enhancement coefficient of the aerobic exercise knowledge node is 0.7, which is greater than the threshold, then it will be output as an extended mining node based on the prior focus on improving cardiorespiratory function. Further exploration may be needed to uncover knowledge about suitable aerobic exercise intensity, duration, and frequency for improving the user's cardiorespiratory function. Similarly, if the flexibility training knowledge node has a knowledge enhancement coefficient of 0.6, which is greater than the threshold, it will also become an extended mining node, further exploring knowledge such as the specificity of different yoga postures in improving flexibility and the optimal practice time. However, if the strength training knowledge node has a knowledge enhancement coefficient of 0.4, which is less than the threshold, it will not be output as an extended mining node.

[0110] In the field of Traditional Chinese Medicine (TCM) emotional well-being, tracking parameters may include emotion type (such as the frequency and intensity of anxiety and depression) and coping strategies (such as talking to others and self-regulation). The prior concern trend hypothesis is to reduce anxiety and improve psychological resilience. Knowledge nodes may include meditation, deep breathing, and social activities. For the meditation knowledge node, if the user's anxiety is severe and meditation is considered an effective way to relieve anxiety, its knowledge reinforcement coefficient may be high. For the deep breathing knowledge node, if it helps relieve anxiety to some extent but with a relatively weak effect, its knowledge reinforcement coefficient may be at a moderate level. For the social activity knowledge node, if it has a positive effect on reducing anxiety and improving psychological resilience and is consistent with the user's situation (such as the user having few social activities and anxiety), its knowledge reinforcement coefficient may be high. A threshold value of 0.55 is set, and the knowledge reinforcement coefficient for the meditation knowledge node is 0.7. If the value is greater than the threshold, it will be output as an extended mining node on the prior concern trend of reducing anxiety. Further mining may then be conducted to explore knowledge such as the differences in the effects of different meditation methods (such as mindfulness meditation and transcendental meditation) on user anxiety relief, and the optimal meditation duration. Social activity knowledge nodes, if their knowledge enhancement coefficient is 0.6 (greater than the threshold), will become extended mining nodes. Further exploration will be conducted to uncover information such as which social activities (e.g., interest groups, community events) are most effective in reducing anxiety and improving stress resistance. Conversely, deep breathing knowledge nodes, if their knowledge enhancement coefficient is 0.5 (less than the threshold), will not be output as extended mining nodes. In this way, the server can filter out knowledge nodes from a large pool that are worthy of focused attention and in-depth exploration under various prior trends, providing a crucial foundation for subsequent optimization of health knowledge and the construction of health plans.

[0111] In one possible implementation, step S130 further includes:

[0112] Step S133: Encode the multiple health feedback data to generate a first encoded representation vector for each of the extended mining nodes.

[0113] Step S134: Obtain the first encoded representation vector of multiple extended mining nodes on each prior interest trend and the interest mapping factor of each extended mining node.

[0114] Step S135: Determine the fusion calculation result between the interest mapping factor of each of the extended mining nodes and the first encoded representation vector.

[0115] Step S136: Based on the prior attention trend, interest mapping is performed on the fusion calculation results, and the fusion calculation results of each extended mining node are accumulated to generate an interest mapping vector that maps to the TCM health preservation knowledge profile of the target user.

[0116] In this embodiment, firstly, the server encodes multiple health feedback data to generate a first encoded representation vector for each extended mining node. Taking the field of traditional Chinese medicine diet and health preservation as an example, assume that the previously determined extended mining nodes include oily food and high-fiber food. The health feedback data includes the user's blood lipid level, weight change, dietary habits (such as the type and amount of food at each meal), and some related physical symptoms (such as indigestion). For the oily food extended mining node, the server encodes these health feedback data. For example, blood lipid levels are encoded according to different ranges, such as 0 for normal range, 1 for slightly high, and 2 for moderately high; weight change is encoded as 1 if it increases, -1 if it decreases, and 0 if it remains unchanged; high frequency of oily food intake in dietary habits is encoded as 1, and low frequency as 0; the presence of indigestion symptoms is encoded as 1, and the absence of indigestion symptoms is encoded as 0. Combining these codes forms the first encoded representation vector for the oily food extended mining node, which may be [1, 1, 1, 1] (this is just a simple example; the actual encoding will be more complex). For the high-fiber food extended mining node, the health feedback data is encoded according to similar rules. Assuming the first encoded vector is [-1, 0, 0, 0], because high-fiber foods help lower blood lipids (if blood lipids are high, it is encoded as -1 to indicate an improvement), have little relation to weight gain (encoded as 0), users may consume less (encoded as 0), and have little relation to indigestion (encoded as 0).

[0117] Next, the server obtains the first encoded representation vector of multiple extended mining nodes on each prior interest trend and the interest mapping factor of each extended mining node. Taking improving blood lipid status as the prior interest trend in the field of traditional Chinese medicine diet and health preservation as an example, for the oily food extended mining node, assuming its interest mapping factor is 0.8 (this value was previously determined based on factors such as the deviation between the health preservation association vector and the prior interest trend), its first encoded representation vector is [1, 1, 1, 1]; for the high fiber food extended mining node, its interest mapping factor is 0.6, and its first encoded representation vector is [-1, 0, 0, 0].

[0118] Then, the server determines the fusion calculation result between the interest mapping factor and the first encoded representation vector for each extended mining node. For the oily food extended mining node, assuming a weighted multiplication fusion calculation method is used (in reality, it may be a more complex calculation method), the interest mapping factor 0.8 is multiplied by each element in the first encoded representation vector [1, 1, 1, 1], resulting in a fusion calculation result of [0.8, 0.8, 0.8, 0.8]. For the high-fiber food extended mining node, the interest mapping factor 0.6 is multiplied by the first encoded representation vector [-1, 0, 0, 0], resulting in a fusion calculation result of [-0.6, 0, 0, 0].

[0119] Finally, based on prior interest trends, the server performs interest mapping on the fusion calculation results, accumulating the fusion calculation results of each extended mining node to generate an interest mapping vector that maps to the target user's TCM health preservation knowledge profile. Under the prior interest trend of improving blood lipid levels, the server accumulates the fusion calculation results of the extended mining nodes for oily foods and high-fiber foods. This accumulation method may involve adding corresponding elements; for example, adding [0.8, 0.8, 0.8, 0.8] for oily foods and [-0.6, 0, 0, 0] for high-fiber foods yields [0.2, 0.8, 0.8, 0.8]. This result is the interest mapping vector mapped to the target user's TCM health preservation knowledge profile. This vector reflects the combined impact of these extended mining nodes on the user's health preservation status under the prior interest trend of improving blood lipid levels.

[0120] Looking at the field of TCM exercise and health preservation, assuming the prior focus is on improving cardiopulmonary function, the expanded mining nodes include aerobic exercise and flexibility training. Health feedback data includes users' cardiopulmonary function indicators (such as vital capacity, heart rate recovery time, etc.), exercise habits (such as weekly exercise frequency, duration of each exercise session, etc.), and feelings of fatigue. For the aerobic exercise expanded mining node, the health feedback data is encoded to obtain a first encoded representation vector, such as [1, 1, 1] (assuming low vital capacity is encoded as 1 indicating a need for improvement, low exercise frequency as 1, and strong post-exercise fatigue as 1), with an interest mapping factor of 0.7. For the flexibility training expanded mining node, the encoded first encoded representation vector is [0, 0, 0] (because flexibility training has little impact on cardiopulmonary function indicators, and exercise habits and fatigue are not closely related to it), with an interest mapping factor of 0.3. After fusion calculation, aerobic exercise yields [0.7, 0.7, 0.7], and flexibility training yields [0, 0, 0]. The accumulated values ​​[0.7, 0.7, 0.7] are used as an interest mapping vector that maps to the target user's profile of TCM health preservation knowledge.

[0121] In the field of Traditional Chinese Medicine (TCM) emotional well-being, the prior focus trend is on reducing anxiety, with extended mining nodes including meditation and social activities. Health feedback data includes anxiety level scores, social frequency, and emotional stability. For the meditation extended mining node, the first encoded vector is [1, 0, 1] (assuming high anxiety level is encoded as 1, social frequency has little relation to meditation as 0, and emotional instability as 1), with an interest mapping factor of 0.8. For the social activity extended mining node, the first encoded vector is [-1, 1, 0] (social activities help reduce anxiety as -1, high social frequency is encoded as 1, and little relation to emotional stability is encoded as 0), with an interest mapping factor of 0.6. After fusion calculation, meditation yields [0.8, 0, 0.8], and social activities yield [-0.6, 0.6, 0]. Accumulated, this results in [0.2, 0.6, 0.8], which serves as the interest mapping vector mapped to the target user's TCM well-being knowledge profile. Through this interest mapping process, the server can more accurately grasp the relationship between the expanded mining nodes and the health needs of target users, providing an important basis for subsequent optimization of the health interest graph and construction of health solutions.

[0122] In one possible implementation, step S140 includes:

[0123] Step S141: Obtain the interest mapping vector and actual interest vector of the extended mining node in the TCM health preservation knowledge profile of the target user.

[0124] Step S142: Determine the feature distance between the interest mapping vector and the actual interest vector.

[0125] Step S143: Determine the error parameters of the TCM health preservation knowledge association model based on the feature distance. The TCM health preservation knowledge association model is used to predict each first association degree in the first health preservation interest graph.

[0126] Step S144: Optimize the model parameter information of the TCM health preservation knowledge association model based on the error parameters to generate an optimized TCM health preservation knowledge association model.

[0127] Step S145: Use the optimized TCM health preservation knowledge association model to iteratively optimize the first health preservation interest graph of the target user, and generate an optimized health preservation interest graph in the TCM health preservation knowledge field of the target user.

[0128] Step S150 includes:

[0129] Step S151: Analyze the optimized health and wellness interest graph, identify various knowledge nodes and the relationships between them, and generate a basic framework for a health and wellness plan that includes multiple health and wellness modules.

[0130] Step S152: For each health preservation section, analyze the depth and breadth of health preservation information contained in the corresponding knowledge nodes in the optimized health preservation interest graph, obtain knowledge node information, and make preliminary plans for each health preservation section based on the knowledge node information, generate preliminary planning content for each health preservation section, and integrate the preliminary planning content of each health preservation section to form a complete TCM health preservation plan.

[0131] In this embodiment, firstly, the server obtains the interest mapping vector and actual interest vector of the extended mining nodes in the target user's TCM health knowledge profile. Taking the TCM dietary health field as an example, the previously determined extended mining nodes, such as oily food and high-fiber food, have corresponding interest mapping vectors. Assume the interest mapping vector for the oily food extended mining node is [0.2, 0.8, 0.8, 0.8] (this vector is derived after comprehensively considering various factors, representing the mapping of oily food in different health aspects). The actual interest vector is obtained through analysis of the user's actual behavior and preferences. For example, by analyzing the user's dietary records, it is found that the user actually frequently consumes oily food, even though oily food may have many disadvantages from a health perspective. Therefore, the actual interest vector might be [1, 1, 1, 1] (representing a high actual inclination towards oily food in terms of blood lipids, weight, dietary habits, and digestion). For the high-fiber food extension mining node, the interest mapping vector is assumed to be [-0.6, 0, 0, 0], while the actual interest vector may be [0, 0, 0, 0], because users may rarely eat high-fiber foods and have low actual interest in them in various health-related aspects.

[0132] Next, the server determines the feature distance between the interest mapping vector and the actual interest vector. This feature distance can be calculated in various ways, such as using the Euclidean distance method. For the oily food extension mining node, the interest mapping vector [0.2, 0.8, 0.8, 0.8] and the actual interest vector [1, 1, 1, 1] are substituted into the Euclidean distance formula to calculate the distance between them. This distance reflects the degree of difference between theoretical health advice (interest mapping vector) and the user's actual behavior (actual interest vector). For the high-fiber food extension mining node, the feature distance between its interest mapping vector [-0.6, 0, 0, 0] and the actual interest vector [0, 0, 0, 0] is also calculated.

[0133] Then, the error parameter of the TCM health preservation knowledge association model is determined based on the feature distance. The TCM health preservation knowledge association model is an important tool for predicting each first association degree in the first health preservation interest graph. Since the feature distance reflects the difference between the interest mapping vector and the actual interest vector, the larger this difference, the greater the potential error in the model's prediction of association degrees. For example, if the feature distance of the oily food extension mining node is large, it means that the TCM health preservation knowledge association model may have a large deviation in predicting the association degree between oily food and the target user's TCM health preservation knowledge profile, thus determining a correspondingly large error parameter based on this feature distance. For the high-fiber food extension mining node, if the feature distance is small, the corresponding error parameter will also be small.

[0134] Next, the model parameters of the TCM health preservation knowledge association model are optimized based on the error parameters to generate an optimized TCM health preservation knowledge association model. It is assumed that the TCM health preservation knowledge association model includes parameters such as the nutritional components of various foods, their effects on the body, their relationship with different body constitutions, and the weight relationships between these parameters. If the error parameter corresponding to oily foods is large, the server may adjust the parameters related to oily foods, such as adjusting the weight parameter of the impact of high fat content of oily foods on blood lipids, or adjusting the parameters related to the relationship between oily foods and spleen and stomach function. For high-fiber foods, if the error parameter is small, only minor adjustments or no adjustment to their related parameters may be made. Through such adjustments, the optimized TCM health preservation knowledge association model is generated.

[0135] Finally, the optimized TCM health preservation knowledge association model is used to iteratively optimize the target user's primary health preservation interest graph, generating an optimized health preservation interest graph within the target user's TCM health preservation knowledge domain. In this process, the optimized health preservation interest graph is analyzed to identify various knowledge nodes and the relationships between them, generating a basic framework for a health preservation plan encompassing multiple health preservation modules. For example, within the TCM health preservation knowledge domain, modules such as dietary health preservation, exercise health preservation, and emotional health preservation are identified. In the dietary health preservation module, knowledge nodes may include various foods (such as the previously mentioned oily foods and high-fiber foods), dietary patterns, etc.; in the exercise health preservation module, knowledge nodes may include various exercise methods (such as running and yoga), exercise intensity, etc.; in the emotional health preservation module, knowledge nodes include methods of emotion regulation (such as meditation and social activities), etc., and the relationships between these knowledge nodes are clarified, such as the mutual influence between diet and emotions (poor diet may affect mood, and mood may also affect food choices), and the correlation between exercise and diet in terms of energy metabolism, thereby constructing the basic framework for a health preservation plan.

[0136] For each health and wellness section, the depth and breadth of health and wellness information contained in the corresponding knowledge nodes in the health and wellness interest graph are analyzed and optimized. This yields knowledge node information, and based on this information, a preliminary plan is developed for each health and wellness section, generating preliminary content for each section. These preliminary plans are then integrated to form a complete TCM health and wellness plan. In the dietary health and wellness section, for the knowledge node on oily foods, the depth of its health and wellness information in the optimized health and wellness interest graph is analyzed. This includes a deeper understanding of the different types of oily foods (the difference between animal and vegetable oils), the degree of oiliness under different cooking methods, and the differences in their health effects. In terms of breadth, the relationship between oily foods and other health and wellness factors is considered, such as their relationship with season, region, and individual constitution. Based on this knowledge node information, the content of the dietary health and wellness section is preliminarily planned, such as recommending a reduction in the intake of oily foods, especially animal oils, which should be strictly controlled in summer and for individuals with a hot and humid constitution. For the exercise and wellness section, the analysis focuses on knowledge points related to exercise methods. For example, the depth of information on the wellness benefits of running includes the impact of different running speeds and distances on cardiopulmonary function and muscle strength, while the breadth covers the synergistic effects of running with other wellness practices (such as diet and emotional well-being). Based on this information, the content of the exercise and wellness section is initially planned, such as suggesting that users choose appropriate running intensity and distance based on their individual physical condition and wellness goals, and paying attention to dietary adjustments before and after exercise. For the emotional wellness section, the analysis focuses on knowledge points related to emotion regulation methods. For example, the depth of information on meditation includes the differences in the effects of different meditation schools and techniques on emotion regulation, while the breadth covers the interaction between meditation and lifestyle (such as diet and exercise). Based on these initial plans, the content of each wellness section is then integrated to form a complete Traditional Chinese Medicine (TCM) wellness plan, such as comprehensive suggestions on diet, exercise, and emotions, providing users with comprehensive TCM wellness guidance.

[0137] Figure 2 This illustration shows the hardware structure of a big data-based personalized recommendation system for TCM health preservation plans 100, provided by an embodiment of the present invention, for implementing the above-described method for personalized recommendation of TCM health preservation plans based on big data. Figure 2 As shown, the personalized recommendation system for TCM health preservation based on big data 100 may include a processor 110, a machine-readable storage medium 120, a bus 130, and a communication unit 140.

[0138] Machine-readable storage medium 120 may store data and / or instructions. In some embodiments, machine-readable storage medium 120 may store data acquired from an external terminal. In some embodiments, machine-readable storage medium 120 may store data and / or instructions used by the big data-based personalized recommendation system for TCM health preservation plans 100 to perform or use in order to complete the exemplary methods described in this invention.

[0139] In the specific implementation process, one or more processors 110 execute computer-executable instructions stored in machine-readable storage medium 120, so that processor 110 can execute the personalized recommendation method of TCM health preservation plan based on big data as described in the above method embodiment. Processor 110, machine-readable storage medium 120 and communication unit 140 are connected through bus 130. Processor 110 can be used to control the sending and receiving actions of communication unit 140.

[0140] The specific implementation process of processor 110 can be found in the various method embodiments executed by the above-mentioned big data-based personalized recommendation system for TCM health preservation, which have similar implementation principles and technical effects. This embodiment will not be repeated here.

[0141] Furthermore, this embodiment of the invention also provides a readable storage medium, wherein computer-executable instructions are preset in the readable storage medium, and when the processor executes the computer-executable instructions, the above-mentioned personalized recommendation method for TCM health preservation based on big data is implemented.

[0142] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for personalized recommendation of TCM health preservation plans based on big data, characterized in that, The method includes: Based on multiple health feedback data of the target user and the tracking parameters corresponding to each health feedback data, a first health interest graph in the TCM health knowledge field of the target user is constructed; the first health interest graph includes the first correlation degree between each knowledge node in the TCM health knowledge field and the TCM health knowledge profile of the target user; Based on the first relevance of each knowledge node and the predefined prior attention trend, the knowledge reinforcement coefficient and interest mapping factor of the corresponding knowledge node are determined respectively; Based on the knowledge enhancement coefficient, all knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing to generate multiple extended mining nodes. Based on the interest mapping factor, interest mapping is performed on the multiple extended mining nodes to generate an interest mapping vector for each extended mining node. Based on the interest mapping vector of the extended mining node, the first health care interest graph of the target user is iteratively optimized to generate an optimized health care interest graph in the field of traditional Chinese medicine health care knowledge of the target user. Based on the optimized health and wellness interest graph, a traditional Chinese medicine (TCM) health and wellness plan is constructed for the target user, and a TCM health and wellness plan recommended to the target user is generated. The step of determining the knowledge reinforcement coefficient and interest mapping factor for each knowledge node based on the first relevance degree and the predefined prior attention trend includes: Based on the first degree of relevance of each knowledge node, determine the knowledge reinforcement coefficient of the corresponding knowledge node; Based on the first relevance of each knowledge node and the predefined prior attention trend, the interest mapping factor of the corresponding knowledge node is determined; The step of determining the knowledge reinforcement coefficient of the corresponding knowledge node based on the first relevance of each knowledge node includes: Obtain predefined knowledge reinforcement functions; The knowledge clustering degree of knowledge reinforcement processing in this round of iterative optimization is determined; the knowledge clustering degree is a feature value determined based on the prior optimized health and wellness interest graph, which is the optimized health and wellness interest graph generated in the previous round of iterative optimization before this round of iterative optimization; the knowledge clustering degree includes a first knowledge clustering degree and a second knowledge clustering degree, wherein the first knowledge clustering degree is greater than the second knowledge clustering degree; when iteratively optimizing the first health and wellness interest graph, the knowledge reinforcement coefficient is iteratively determined based on the first knowledge clustering degree and the second knowledge clustering degree. The knowledge aggregation degree and the first correlation degree of each knowledge node are loaded into the knowledge reinforcement function to generate the knowledge reinforcement coefficient of the corresponding knowledge node; The step of determining the interest mapping factor for each knowledge node based on the first relevance degree of each knowledge node and the predefined prior attention trend includes: Obtain the health preservation association vector that maps each knowledge node to the TCM health preservation knowledge profile node of the target user, and determine the deviation value between the health preservation association vector and the prior attention trend; Obtain a predefined interest mapping function and determine candidate adjustment parameters for the interest mapping in the current iteration of optimization. The candidate adjustment parameters are feature values ​​determined based on the prior optimized health and wellness interest map, which is the optimized health and wellness interest map generated in the previous iteration of optimization before the current iteration of optimization. The candidate adjustment parameters and the deviation values ​​are loaded into the interest mapping function to generate the interest mapping factors for the corresponding knowledge nodes.

2. The method for personalized recommendation of TCM health preservation plans based on big data according to claim 1, characterized in that, The step of constructing a first health interest map in the field of traditional Chinese medicine health knowledge of the target user based on multiple health feedback data of the target user and tracking parameters corresponding to each health feedback data includes: Extract the health preservation scope information and health preservation monitoring parameters of the TCM health preservation knowledge field from the tracking parameters; The target user's area of ​​TCM health preservation knowledge is determined based on the aforementioned health preservation scope information; Based on the aforementioned health monitoring parameters, determine the attribute information of each knowledge node in the domain of traditional Chinese medicine health preservation knowledge; The attribute information and the multiple health feedback data are loaded into the TCM health preservation knowledge association model to generate the first association degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user; By utilizing the first correlation degree between each knowledge node in the TCM health preservation knowledge domain and the TCM health preservation knowledge profile of the target user, a first health preservation interest map of the TCM health preservation knowledge domain in which the target user resides is constructed.

3. The method for personalized recommendation of TCM health preservation plans based on big data according to claim 1, characterized in that, Based on the knowledge enhancement coefficient, knowledge enhancement processing is performed on all knowledge nodes in the TCM health preservation knowledge domain to generate multiple extended mining nodes, including: Based on the knowledge enhancement coefficient, the knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing in the prior attention trend corresponding to the tracking parameters, thereby generating the multiple extended mining nodes.

4. The method for personalized recommendation of TCM health preservation plans based on big data according to claim 3, characterized in that, Based on the knowledge enhancement coefficient, the knowledge nodes in the TCM health preservation knowledge domain are subjected to knowledge enhancement processing in the prior attention trend corresponding to the tracking parameters to generate the multiple extended mining nodes, including: For each of the multiple prior attention trends corresponding to the tracking parameters, obtain the knowledge reinforcement coefficient of each of the multiple knowledge nodes. Knowledge nodes whose knowledge enhancement coefficient is greater than the threshold value are output as extended mining nodes on the prior attention trend.

5. The method for personalized recommendation of TCM health preservation plans based on big data according to claim 4, characterized in that, The step of performing interest mapping on the plurality of extended mining nodes based on the interest mapping factor, and generating an interest mapping vector for each of the extended mining nodes, includes: The multiple health feedback data are encoded and represented to generate a first encoded representation vector for each of the extended mining nodes; Obtain the first encoded representation vector of multiple extended mining nodes on each prior interest trend and the interest mapping factor of each extended mining node; Determine the fusion calculation result between the interest mapping factor of each of the extended mining nodes and the first encoded representation vector; Based on the prior attention trend, the fusion calculation results are mapped to interests, and the fusion calculation results of each extended mining node are accumulated to generate an interest mapping vector that maps to the TCM health preservation knowledge profile of the target user.

6. The method for personalized recommendation of TCM health preservation plans based on big data according to any one of claims 1-5, characterized in that, The step of iteratively optimizing the first health and wellness interest graph of the target user based on the interest mapping vector of the extended mining node to generate an optimized health and wellness interest graph in the field of traditional Chinese medicine health and wellness knowledge of the target user includes: Obtain the interest mapping vector and actual interest vector of the extended mining node in the target user's TCM health preservation knowledge profile; Determine the feature distance between the interest mapping vector and the actual interest vector; The error parameters of the TCM health preservation knowledge association model are determined based on the feature distance; the TCM health preservation knowledge association model is used to predict each first association degree in the first health preservation interest graph. Based on the error parameters, the model parameter information of the TCM health preservation knowledge association model is optimized to generate an optimized TCM health preservation knowledge association model; The optimized TCM health preservation knowledge association model is used to iteratively optimize the first health preservation interest graph of the target user, thereby generating an optimized health preservation interest graph in the TCM health preservation knowledge domain of the target user. The step of constructing a TCM health preservation plan for the target user based on the optimized health preservation interest graph and generating a TCM health preservation plan recommended to the target user includes: The optimized health and wellness interest graph is analyzed to identify various knowledge nodes and the relationships between them, generating a basic framework for a health and wellness plan that includes multiple health and wellness modules. For each health preservation section, the depth and breadth of health preservation information contained in the corresponding knowledge nodes in the optimized health preservation interest graph are analyzed to obtain knowledge node information. Based on the knowledge node information, a preliminary plan is made for each health preservation section, generating the preliminary plan content for each health preservation section. The preliminary plan content of each health preservation section is integrated to form a complete TCM health preservation plan.

7. A personalized recommendation system for TCM health preservation plans based on big data, characterized in that, The big data-based personalized TCM health preservation plan recommendation system includes a processor and a memory, the memory and the processor being connected. The memory is used to store programs, instructions or code, and the processor is used to execute the programs, instructions or code in the memory to implement the big data-based personalized TCM health preservation plan recommendation method as described in any one of claims 1-6.

Images