A traditional Chinese medicine diet therapy intervention pushing method and device based on multi-source data and a storage medium

Abstract

The application provides a traditional Chinese medicine diet therapy intervention pushing method and device based on multi-source data and a storage medium, and relates to the technical field of intelligent diet therapy. The method periodically collects initial health state data of a user and obtains first health feature data through standardization, generates second health feature data through feature extraction and fusion, matches traditional Chinese medicine constitution rules to obtain a constitution label, generates a candidate diet therapy scheme through reasoning of a diet therapy knowledge graph based on the label, obtains an optimized scheme through multi-dimensional evaluation and sorting, and pushes the optimized scheme to a user terminal. The application dynamically captures body changes through periodic collection, eliminates the lag of single static data recommendation, unifies data formats through standardization, provides a standard basis for quantitative operation, integrates multi-dimensional health information through feature fusion, digitally and objectively determines the constitution of a user, generates adaptive diet therapy combinations in batches with the help of a knowledge graph combined with drug properties and contraindications, and realizes digital output of diet therapy through linkage of each link.

Description

Technical Field

[0001] This invention mainly relates to the field of smart diet therapy technology, specifically to a method, device, and storage medium for pushing traditional Chinese medicine diet therapy interventions based on multi-source data. Background Technology

[0002] In current public health management scenarios, traditional Chinese medicine (TCM) dietary therapy has become a mainstream choice for daily health maintenance. Traditional dietary therapy recommendations largely rely on in-person consultations with TCM doctors: doctors ask users about their physical sensations and lifestyle habits face-to-face, subjectively assess their constitution based on basic physical signs, and then verbally or in writing recommend suitable ingredients and medicinal diets. This process depends entirely on the doctor's personal experience, with extremely low levels of standardization and digitization. With the popularization of online health services, a number of simple dietary therapy recommendation tools have emerged. These tools merely store a standardized list of medicinal diets and rely solely on users manually selecting single symptoms for simple matching. They can only perform rough keyword filtering and cannot integrate multi-dimensional information such as long-term physiological indicators, seasonal and regional factors, dietary restrictions, and dynamic changes in constitution. They lack complete data flow and quantitative calculation logic.

[0003] Existing technologies lack both the ability to standardize and process multi-source heterogeneous health data and the ability to build a quantitative constitution identification model for objective constitution assessment. Furthermore, they lack a knowledge graph reasoning mechanism based on the properties and medicinal theories of traditional Chinese medicine, only able to output fixed, template-based dietary therapy content, unable to differentiate and rank the effectiveness, safety, and individual suitability of the plans. In addition, existing push notification models lack dynamic timing control logic, with completely uniform push content and time periods. They cannot iteratively update dietary therapy plans based on user terminal usage habits and fluctuations in constitution, resulting in highly homogenized and untargeted content. This easily leads to problems such as food incompatibility with user constitution, and neglect of user allergies and past intolerances. The precision of dietary therapy intervention is insufficient, failing to meet the actual needs of different groups for personalized and continuous traditional Chinese medicine dietary therapy. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a method, device and storage medium for pushing traditional Chinese medicine dietary therapy intervention based on multi-source data, which addresses the shortcomings of the prior art.

[0005] The technical solution of this invention to solve the above-mentioned technical problems is as follows: A method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, comprising the following steps: According to a preset time period, the initial health status data of the target user is periodically collected, and the initial health status data is standardized to obtain the first health feature data. The first health feature data is subjected to feature extraction and fusion processing to generate the second health feature data; The second health feature data is matched with the preset TCM constitution classification rules to obtain the constitution type label data of the target user; Based on the body constitution type label data, association retrieval and reasoning are performed in the preset TCM diet therapy knowledge graph to generate candidate diet therapy intervention plan data; The candidate dietary intervention data are evaluated for suitability and ranked to obtain the optimal dietary intervention data; The data of the preferred dietary therapy intervention plan is pushed to the target user's terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

[0006] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: A device for pushing traditional Chinese medicine dietary therapy intervention based on multi-source data, comprising: The data acquisition and standardization module is used to periodically collect the initial health status data of the target user according to a preset time period, and to standardize the initial health status data to obtain the first health feature data. The feature extraction and fusion module is used to perform feature extraction and fusion processing on the first health feature data to generate the second health feature data. The TCM constitution identification module is used to match the second health feature data with the preset TCM constitution classification rules to obtain the constitution type label data of the target user. The dietary therapy plan reasoning module is used to perform association retrieval and reasoning processing in a preset traditional Chinese medicine dietary therapy knowledge graph based on the constitution type label data, and generate candidate dietary therapy intervention plan data. The scheme evaluation and optimization module is used to evaluate and rank the candidate dietary intervention scheme data to obtain the optimized dietary intervention scheme data. The personalized push module is used to push the data of the preferred dietary therapy intervention plan to the target user terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

[0007] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a traditional Chinese medicine dietary therapy intervention push device based on multi-source data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it realizes the traditional Chinese medicine dietary therapy intervention push method based on multi-source data as described above.

[0008] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows: a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-described method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data.

[0009] The beneficial effects of this invention are as follows: By continuously collecting users' initial health status data at a preset cycle, changes in the body can be dynamically captured, solving the problem of lag in single static data recommendations; the raw data is standardized and processed to unify the data expression format, providing a standardized data source for subsequent quantitative calculations; feature extraction and fusion are performed on the standardized data to integrate multi-dimensional information into complete characteristic data representing the user's health; the fused features are matched with constitution rules to generate constitution tags, objectively determining the user's unbalanced constitution through digital means; based on the constitution tags, the invention uses a dietary therapy knowledge graph for retrieval and reasoning, and combines the properties and contraindications of traditional Chinese medicine to generate suitable dietary therapy combinations in batches; candidate solutions are evaluated and ranked in multiple dimensions to screen out high-risk and low-suitability dietary therapy content; finally, the preferred solution is pushed to the terminal, delivering customized conditioning content. Each step is linked sequentially to complete the personalized digital output of dietary therapy, effectively improving the targeting and timeliness of dietary therapy recommendations. Attached Figure Description

[0010] Figure 1 A flowchart of the method for pushing traditional Chinese medicine dietary therapy interventions provided in an embodiment of the present invention; Figure 2 This is a block diagram of a traditional Chinese medicine dietary therapy intervention push device provided in an embodiment of the present invention. Detailed Implementation

[0011] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0012] Example 1: As Figure 1 As shown, this embodiment of the invention provides a method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, including the following steps: S1. Collect the initial health status data of the target user periodically according to the preset time period, and standardize the initial health status data to obtain the first health feature data. S2. Perform feature extraction and fusion processing on the first health feature data to generate second health feature data; S3. Match the second health feature data with the preset TCM constitution classification rules to obtain the constitution type label data of the target user; S4. Based on the body constitution type tag data, perform association retrieval and reasoning processing in the preset traditional Chinese medicine diet therapy knowledge graph to generate candidate diet therapy intervention plan data; S5. Perform an adaptation assessment and sorting process on the candidate dietary intervention data to obtain the optimal dietary intervention data; S6. Push the data of the preferred dietary therapy intervention plan to the target user terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

[0013] In the above embodiments, by continuously collecting the initial health status data of the target user according to a preset cycle, the changes in the user's body at different stages can be captured in real time, avoiding the lag caused by relying on single static health data for dietary therapy recommendations; by standardizing the collected initial health status data, the expression form of various health data can be unified, providing a regular data foundation for subsequent quantitative calculations; by extracting and fusing features from the standardized data, multi-dimensional health information can be integrated to form feature data that can completely represent the user's physical condition; by matching the fused feature data with the TCM constitution classification rules to obtain constitution tags, the user's unbalanced constitution can be objectively determined digitally; by generating candidate dietary therapy plans by searching and reasoning in the dietary therapy knowledge graph based on constitution tags, suitable food combinations can be generated in batches by combining the properties of TCM medicines and the logic of what is suitable and unsuitable; by conducting multi-dimensional adaptation evaluation and sorting of candidate dietary therapy plans, dietary therapy content with high safety risks and low matching degree can be eliminated; by pushing the preferred dietary therapy plan to the user terminal, customized conditioning content can be directly delivered to the user, completing the digital output of personalized dietary therapy intervention layer by layer, and improving the overall pertinence and timeliness of dietary therapy recommendations.

[0014] Preferably, the initial health status data is standardized to obtain first health characteristic data, including: The initial health status data is divided into structured fields and unstructured fields. The structured fields are used to represent the objective physiological indicator data of the target user, and the unstructured fields are used to represent the subjective health description data of the target user. The structured fields are subjected to dimensional unification processing, and the unstructured fields are subjected to semantic encoding processing to obtain the processed first health feature data.

[0015] In practical implementation, the dimensional unification processing can adopt min-max normalization or Z-score standardization to map physiological indicators of different magnitudes to a unified numerical range, avoiding individual indicators from dominating subsequent calculation results due to excessively large values; semantic encoding processing can be combined with a Chinese medicine terminology dictionary to standardize and encode unstructured descriptions such as "fatigue", "aversion to cold", and "dry mouth", transforming vague subjective feelings into quantifiable features, thereby improving the stability and generalization ability of subsequent constitution identification.

[0016] In the above embodiments, by dividing the initial health status data into structured fields representing physiological indicators and unstructured fields representing subjective feelings, it is possible to distinguish between two completely different types of health information, preventing objective numerical data from being mixed with textual description data and thus preventing them from being used in unified calculations. By implementing dimensional unification processing on the structured fields, it is possible to eliminate calculation biases caused by differences in the units of different indicators such as height, blood pressure, and sleep duration. By performing semantic encoding processing on the unstructured fields, it is possible to transform the textual form of physical feelings into quantitative information that can participate in numerical calculations, thereby allowing objective physiological data and subjective symptom data to participate in subsequent feature fusion calculations simultaneously, ensuring that the input data is complete and usable.

[0017] Preferably, S2, the first health feature data is subjected to feature extraction and fusion processing to generate second health feature data, including: The structured and unstructured fields in the processed first health feature data are encoded and mapped to obtain multiple numerical feature components. These numerical feature components are then used as input to a weighted fusion function to generate the second health feature data. The computational expression for encoding mapping processing is: in, Indicates the first The raw data for each field is divided into structured fields containing objective physiological indicators and unstructured fields containing subjective health descriptions. This represents an encoding mapping function that matches the field type, used for numerical transformation of objective physiological indicator data or semantic encoding of subjective health description data; Indicates the first Numerical feature components obtained after encoding and mapping of each field; The weighted fusion function is expressed as: , in, Represents the total number of numerical characteristic components. Indicates the first The weight coefficients corresponding to each numerical feature component. This represents the second health feature data obtained after fusion, used to characterize the overall health status of the target user.

[0018] In practical implementation, the weighting coefficient The model can be pre-set based on the strength of the influence of various indicators on physical constitution in traditional Chinese medicine theory, or it can be trained by machine learning model based on historical labeled samples, thereby achieving dual optimization of "traditional Chinese medicine experience + data-driven". In addition, a non-linear activation function can be introduced to map the weighted fusion results, enhance the model's ability to express complex health states, and avoid the loss of feature information caused by linear superposition.

[0019] In the above embodiments, by using an encoding mapping function adapted to the field type to process structured and unstructured data respectively, targeted numerical conversion and semantic digitization can be completed to generate numerical feature components with unified standards. By setting a weighted fusion function to sum all numerical feature components to obtain the second health feature data, corresponding weights can be assigned according to the magnitude of the influence of different indicators on physical fitness, highlighting the role of key health information and weakening interfering secondary data, thereby completely and objectively quantifying the user's comprehensive health status and improving the comprehensiveness of subsequent physical fitness identification results.

[0020] Preferably, in step S3, the second health feature data is matched with preset TCM constitution classification rules to obtain the target user's constitution type label data, including: Calculate the matching degree between the second health feature data and the preset TCM constitution standard features, and determine the constitution category with the highest matching degree as the constitution type label data of the target user. The expression for calculating the matching degree is: , in, Indicates the first Traditional Chinese medicine constitution types , This represents the total number of body type categories. This represents the second health characteristic data. Indicates the first The prior probability of a particular constitution type is determined by historical statistical data from traditional Chinese medicine diagnosis and treatment. This indicates that the second health characteristic data originates from the first... Under conditions similar to traditional Chinese medicine constitution types, the probability of this second health characteristic data appearing is... This represents the marginal probability of the second health feature data, used to normalize the posterior probabilities for each body type, so that the sum of the posterior probabilities for all body types is 1. This indicates that, given the second health characteristic data, the target user belongs to the [missing information]. The posterior probability of the type of constitution is used as the matching degree.

[0021] In practical implementation, prior probability It can be dynamically adjusted according to seasons, regions, or population characteristics. For example, in seasons where dampness is more prevalent than heat, the prior weights of "phlegm-dampness constitution" and "damp-heat constitution" can be appropriately increased, so that the constitution determination results are more in line with changes in the external environment. At the same time, a matching degree threshold can be set. When the highest matching degree is lower than the threshold, a manual review process is triggered to avoid misjudgment of constitution due to extreme abnormal data.

[0022] In the above embodiments, the posterior matching degree of the fused features corresponding to various constitutions is solved by the Bayesian probability calculation formula. Historical TCM diagnosis and treatment statistics can be introduced to determine the prior probability of various constitutions, and quantitative probability can replace subjective human judgment. By selecting the constitution category with the highest matching degree as the user's constitution label, a standardized and reproducible constitution judgment result can be formed, avoiding the constitution judgment bias caused by the difference in experience of different doctors, and thus providing a stable and objective basis for constitution judgment for the generation of dietary therapy plans.

[0023] Preferably, S4, based on the constitution type tag data, perform association retrieval and reasoning processing in a preset traditional Chinese medicine dietary therapy knowledge graph to generate candidate dietary therapy intervention plan data, including: Based on the constitution type tag data, food ingredient nodes, efficacy nodes, and contraindication nodes associated with the constitution type are retrieved from the traditional Chinese medicine dietary therapy knowledge graph. Multi-hop reasoning is then performed based on the semantic relationships in the graph to generate several candidate dietary therapy intervention schemes. The traditional Chinese medicine dietary therapy knowledge graph is composed of triples, represented as follows: , in, This represents a set of entities, including physical constitution entities, food ingredients entities, efficacy entities, and contraindication entities. This represents a set of relationships, including at least one of the following: suitable for consumption, possessing medicinal properties, and having dietary contraindications. Represents a de novo entity End-to-end entity Relationship ; The multi-hop reasoning process is represented as follows: , in, This represents the body type label data. Representing continuous semantic relationships in a graph. This represents the intermediate and final entities in the reasoning path. Indicates the first Each reasoning path corresponds to a candidate dietary intervention plan data point. All inference paths that meet the conditions are output as candidate dietary intervention schemes.

[0024] In practical implementation, the multi-hop reasoning can also introduce synergistic and antagonistic relationships between ingredients, such as "Astragalus combined with Angelica can enhance the effect of replenishing qi and nourishing blood" and "Radish counteracts the medicinal effects of ginseng", so as to select ingredient combinations with strong synergy and weak conflict in the reasoning path; in addition, the reasoning path can be pruned by combining the local medicinal materials and dietary habits of the user's region to improve the availability and feasibility of candidate solutions.

[0025] In the above embodiments, by constructing a dietary therapy knowledge graph containing entities such as constitution, ingredients, efficacy, and contraindications in the form of triplets, it is possible to fully store the professional semantic relationships such as suitability, incompatibility, and efficacy association between traditional Chinese medicine ingredients and constitution. By relying on user constitution tags to carry out multi-hop relationship retrieval and reasoning in the graph, it is possible to connect multiple layers of entity associations to mine compound dietary therapy combination schemes, no longer limited to single-layer direct matching of constitution and ingredients. By generating candidate dietary therapy intervention schemes corresponding to all reasoning paths, it is possible to simultaneously filter ingredient combinations with dietary contraindications, thereby expanding the range of dietary therapy options while ensuring basic food safety.

[0026] Preferably, step S5 involves performing a suitability assessment and ranking process on the candidate dietary intervention data to obtain preferred dietary intervention data, including: For each candidate dietary intervention plan, scores were calculated in three dimensions: effectiveness, safety, and user suitability. These multi-dimensional scores were then weighted and fused to obtain a comprehensive suitability score. Candidate dietary intervention plans were ranked from highest to lowest according to their comprehensive suitability scores, and the top-ranked plans were selected as the preferred dietary intervention plans. The comprehensive suitability score for any candidate dietary intervention plan is as follows: , in, Indicates the first The overall suitability score of the candidate dietary intervention programs. express The effectiveness score of each scheme is determined based on the strength of the association between ingredients and efficacy in the aforementioned TCM dietary therapy knowledge graph. Indicates the first The safety score for each scheme is determined based on the degree of conflict between the prohibited entity and the constitution type label data. Indicates the first The user suitability score for each solution is determined based on the target users' historical feedback data and preference characteristics. These are the weight coefficients for the corresponding rating dimensions, and .

[0027] In practical implementation, the weighting coefficient Users can dynamically adjust this according to their own needs; for example, users who prioritize therapeutic effects can increase [their prosody]. The weighting can be increased for users with allergies. The weighting of the scores allows for personalized evaluation preferences; at the same time, historical user feedback data can be incorporated to update the scoring model online, ensuring that the optimal solution is continuously optimized based on actual user experience, thereby enhancing long-term user engagement.

[0028] In the above embodiments, by scoring each candidate dietary therapy plan separately from three independent dimensions—effectiveness of conditioning, safety of consumption, and personal suitability—it is possible to comprehensively consider the three core evaluation criteria of medicinal diet conditioning effects, risk of physical constitution conflict, and user dietary preferences. By setting normalized weight coefficients to weight and fuse the three-dimensional scores to obtain a comprehensive suitability score, the priority of different evaluation dimensions can be flexibly adjusted. By sorting the comprehensive suitability scores in descending order and selecting the top-ranked plans as the preferred plans, dietary therapy content that balances efficacy, safety, and user acceptance can be automatically screened, reducing the problem of mismatch between pushed content and users' actual situation.

[0029] Preferably, step S6 involves pushing the preferred dietary therapy intervention data to the target user's terminal to complete the personalized delivery of traditional Chinese medicine dietary therapy intervention, including: Based on the target user's terminal type and interaction preferences, the display format of the preferred dietary intervention plan data is transformed to generate push content data, which is then sent to the target user's terminal via a push timing function. The push timing function is expressed as follows: , in, Indicates the push notification time. Indicates the time of the last push notification. Indicates the preset push interval. As an indicator function, when the overall fit score Greater than the preset threshold The value is 1 if the condition is met, and 0 otherwise. After the push is completed, the target user's updated health status data is collected again according to the preset time period, and steps S1 to S5 are executed again to iteratively update and continuously push the data of the preferred dietary intervention plan.

[0030] In the above embodiments, by combining the target user's terminal type and interaction preferences to optimize the display format of the dietary therapy plan, it is possible to adapt to the interface logic of different terminals such as mobile phones and tablets, and optimize the user's reading experience; by dynamically adjusting the push timing function, it is possible to avoid pushing at fixed times and disturbing users, thus improving the humanization of the push; by periodically collecting the user's updated health data again after the push is completed and repeating all data processing steps, it is possible to continuously track the dynamic changes in the user's constitution, synchronously iterate and update the dietary therapy intervention plan, and thus realize long-term, dynamic, and continuous personalized TCM dietary therapy conditioning services.

[0031] Example 2: Figure 2 As shown, this embodiment of the invention also provides a device for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, including: The data acquisition and standardization module is used to periodically collect the initial health status data of the target user according to a preset time period, and to standardize the initial health status data to obtain the first health feature data. The feature extraction and fusion module is used to perform feature extraction and fusion processing on the first health feature data to generate the second health feature data. The TCM constitution identification module is used to match the second health feature data with the preset TCM constitution classification rules to obtain the constitution type label data of the target user. The dietary therapy plan reasoning module is used to perform association retrieval and reasoning processing in a preset traditional Chinese medicine dietary therapy knowledge graph based on the constitution type label data, and generate candidate dietary therapy intervention plan data. The scheme evaluation and optimization module is used to evaluate and rank the candidate dietary intervention scheme data to obtain the optimized dietary intervention scheme data. The personalized push module is used to push the data of the preferred dietary therapy intervention plan to the target user terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

[0032] Preferably, the second health feature data is matched with preset TCM constitution classification rules to obtain the constitution type label data of the target user, including: Calculate the matching degree between the second health feature data and the preset TCM constitution standard features, and determine the constitution category with the highest matching degree as the constitution type label data of the target user. The expression for calculating the matching degree is: , in, Indicates the first Traditional Chinese medicine constitution types , This represents the total number of body type categories. This represents the second health characteristic data. Indicates the first The prior probability of a particular constitution type is determined by historical statistical data from traditional Chinese medicine diagnosis and treatment. This indicates that the second health characteristic data originates from the first... Under conditions similar to traditional Chinese medicine constitution types, the probability of this second health characteristic data appearing is... This represents the marginal probability of the second health feature data, used to normalize the posterior probabilities for each body type, so that the sum of the posterior probabilities for all body types is 1. This indicates that, given the second health characteristic data, the target user belongs to the [missing information]. The posterior probability of the type of constitution is used as the matching degree.

[0033] This invention also provides a TCM dietary therapy intervention push device based on multi-source data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the TCM dietary therapy intervention push method based on multi-source data as described above.

[0034] This invention also provides a computer-readable storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data as described above.

[0035] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0036] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the above-described apparatus and unit can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0037] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.

[0038] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0039] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0040] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, characterized in that, Includes the following steps: According to a preset time period, the initial health status data of the target user is periodically collected, and the initial health status data is standardized to obtain the first health feature data. The first health feature data is subjected to feature extraction and fusion processing to generate the second health feature data; The second health feature data is matched with the preset TCM constitution classification rules to obtain the constitution type label data of the target user; Based on the body constitution type label data, association retrieval and reasoning are performed in the preset traditional Chinese medicine diet therapy knowledge graph to generate candidate diet therapy intervention plan data; The candidate dietary intervention data are evaluated for suitability and ranked to obtain the optimal dietary intervention data; The data of the preferred dietary therapy intervention plan is pushed to the target user's terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

2. The method for promoting traditional Chinese medicine dietary therapy intervention according to claim 1, characterized in that, The initial health status data is standardized to obtain first health feature data, including: The initial health status data is divided into structured fields and unstructured fields. The structured fields are used to represent the objective physiological indicator data of the target user, and the unstructured fields are used to represent the subjective health description data of the target user. The structured fields are subjected to dimensional unification processing, and the unstructured fields are subjected to semantic encoding processing to obtain the processed first health feature data.

3. The method for promoting traditional Chinese medicine dietary therapy intervention according to claim 1, characterized in that, S2. Perform feature extraction and fusion processing on the first health feature data to generate second health feature data, including: The structured and unstructured fields in the processed first health feature data are encoded and mapped to obtain multiple numerical feature components. These numerical feature components are then used as input to a weighted fusion function to generate the second health feature data. The computational expression for encoding mapping processing is: in, Indicates the first The raw data for each field is divided into structured fields containing objective physiological indicators and unstructured fields containing subjective health descriptions. This represents an encoding mapping function that matches the field type, used for numerical transformation of objective physiological indicator data or semantic encoding of subjective health description data; Indicates the first Numerical feature components obtained after encoding and mapping of each field; The weighted fusion function is expressed as: ， in, Represents the total number of numerical characteristic components. Indicates the first The weight coefficients corresponding to each numerical feature component This represents the second health feature data obtained after fusion, used to characterize the overall health status of the target user.

4. The method for promoting traditional Chinese medicine dietary therapy intervention according to claim 1, characterized in that, The second health characteristic data is matched with preset TCM constitution classification rules to obtain the target user's constitution type label data, including: Calculate the matching degree between the second health feature data and the preset TCM constitution standard features, and determine the constitution category with the highest matching degree as the constitution type label data of the target user. The expression for calculating the matching degree is: ， in, Indicates the first Traditional Chinese medicine constitution types , This represents the total number of body types. This represents the second health characteristic data. Indicates the first The prior probability of a particular constitution type is determined by historical statistical data from traditional Chinese medicine diagnosis and treatment. This indicates that the second health characteristic data originates from the first... Under conditions similar to traditional Chinese medicine constitution types, the probability of this second health characteristic data appearing is... This represents the marginal probability of the second health feature data, used to normalize the posterior probabilities for each body type, so that the sum of the posterior probabilities for all body types is 1. This indicates that, given the second health characteristic data, the target user belongs to the [missing information]. The posterior probability of the type of constitution is used as the matching degree.

5. The method for promoting traditional Chinese medicine dietary therapy intervention according to claim 1, characterized in that, Based on the aforementioned constitution type tag data, association retrieval and reasoning are performed in a pre-defined traditional Chinese medicine dietary therapy knowledge graph to generate candidate dietary therapy intervention plan data, including: Based on the constitution type tag data, food ingredient nodes, efficacy nodes, and contraindication nodes associated with the constitution type are retrieved from the traditional Chinese medicine dietary therapy knowledge graph. Multi-hop reasoning is then performed based on the semantic relationships in the graph to generate several candidate dietary therapy intervention schemes. The traditional Chinese medicine dietary therapy knowledge graph is composed of triples, represented as follows: ， in, This represents a set of entities, including physical constitution entities, food ingredients entities, efficacy entities, and contraindication entities. This represents a set of relationships, including at least one of the following: suitable for consumption, possessing medicinal properties, and having dietary contraindications. Represents a de novo entity End-to-end entity Relationship ; The multi-hop reasoning process is represented as follows: ， in, This represents the body type label data. Representing continuous semantic relationships in a graph. This represents the intermediate and final entities in the reasoning path. Indicates the first Each reasoning path corresponds to a candidate dietary intervention plan data point. All inference paths that meet the conditions are output as candidate dietary intervention schemes.

6. The method for promoting traditional Chinese medicine dietary therapy intervention according to claim 1, characterized in that, The candidate dietary intervention data were subjected to suitability assessment and ranking to obtain optimized dietary intervention data, including: For each candidate dietary intervention plan, scores were calculated in three dimensions: effectiveness, safety, and user suitability. These multi-dimensional scores were then weighted and fused to obtain a comprehensive suitability score. Candidate dietary intervention plans were ranked from highest to lowest according to their comprehensive suitability scores, and the top-ranked plans were selected as the preferred plans. The comprehensive suitability score for any candidate dietary intervention plan is as follows: ， in, Indicates the first The overall suitability score of the candidate dietary intervention programs. express The effectiveness score of each scheme is determined based on the strength of the association between ingredients and efficacy in the aforementioned TCM dietary therapy knowledge graph. Indicates the first The safety score for each scheme is determined based on the degree of conflict between the prohibited entity and the constitution type label data. Indicates the first The user suitability score for each solution is determined based on the target users' historical feedback data and preference characteristics. These are the weight coefficients for the corresponding rating dimensions, and .

7. A device for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, characterized in that, include: The data acquisition and standardization module is used to periodically collect the initial health status data of the target user according to a preset time period, and to standardize the initial health status data to obtain the first health feature data. The feature extraction and fusion module is used to perform feature extraction and fusion processing on the first health feature data to generate the second health feature data. The TCM constitution identification module is used to match the second health feature data with the preset TCM constitution classification rules to obtain the constitution type label data of the target user. The dietary therapy plan reasoning module is used to perform association retrieval and reasoning processing in a preset traditional Chinese medicine dietary therapy knowledge graph based on the constitution type label data, and generate candidate dietary therapy intervention plan data. The scheme evaluation and optimization module is used to evaluate and rank the candidate dietary intervention scheme data to obtain the optimized dietary intervention scheme data. The personalized push module is used to push the data of the preferred dietary therapy intervention plan to the target user terminal to complete the personalized push of traditional Chinese medicine dietary therapy intervention.

8. The traditional Chinese medicine dietary therapy intervention delivery device according to claim 7, characterized in that, The second health characteristic data is matched with preset TCM constitution classification rules to obtain the target user's constitution type label data, including: Calculate the matching degree between the second health feature data and the preset TCM constitution standard features, and determine the constitution category with the highest matching degree as the constitution type label data of the target user. The expression for calculating the matching degree is: ， in, Indicates the first Traditional Chinese medicine constitution types , This represents the total number of body types. This represents the second health characteristic data. Indicates the first The prior probability of a particular constitution type is determined by historical statistical data from traditional Chinese medicine diagnosis and treatment. This indicates that the second health characteristic data originates from the first... Under conditions similar to traditional Chinese medicine constitution types, the probability of this second health characteristic data appearing is... This represents the marginal probability of the second health feature data, used to normalize the posterior probabilities for each body type, so that the sum of the posterior probabilities for all body types is 1. This indicates that, given the second health characteristic data, the target user belongs to the [missing information]. The posterior probability of the type of constitution is used as the matching degree.

9. A device for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data as described in any one of claims 1 to 6.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for pushing traditional Chinese medicine dietary therapy interventions based on multi-source data as described in any one of claims 1 to 6.

Images