A diet recommendation method and system based on multi-dimensional data analysis

Abstract

This application provides a dietary recommendation method and system based on multi-dimensional data analysis. The method includes: obtaining a set of modern nutritional attributes and a set of traditional Chinese medicine (TCM) theoretical attributes of the target food; obtaining a TCM constitution suitability score by matching the TCM theoretical attribute set with the user's TCM constitution type using a first rule; obtaining a disease nutritional suitability score by matching the modern nutritional attribute set with the user's list of diagnosed diseases using a second rule; determining whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutritional suitability score; if a dietary recommendation conflict exists, assessing the severity of the disease and the deviation of the constitution for the user based on the list of diagnosed diseases and the TCM constitution type, respectively, to obtain corresponding disease severity indicators and constitution deviation indicators; and generating dietary recommendations for the target food based on the disease severity indicators and the constitution deviation indicators, thereby improving the accuracy and rationality of dietary recommendations.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence and health management technology, and in particular to a dietary recommendation method and system based on multi-dimensional data analysis. Background Technology

[0002] With the rapid development of artificial intelligence technology and big data analytics, personalized health management services have become an important development direction in the healthcare field. Currently, dietary recommendation systems based on user health data are widely used in clinical nutrition management and daily health maintenance. These systems typically collect data such as users' disease information, physiological indicators, and dietary preferences, and combine this data with a nutrition knowledge base to generate personalized dietary recommendations for users.

[0003] In recent years, integrative medicine, as an emerging field, has gained unprecedented attention. Its core concept is to combine traditional medical systems with modern evidence-based medicine to provide patients with more comprehensive health management solutions. In the field of dietary health management, traditional Chinese medicine (TCM) dietary therapy and modern nutrition each have their theoretical foundations and practical value. TCM theory, based on a holistic view and the principle of syndrome differentiation and treatment, classifies foods according to their properties, flavors, and meridian tropism, emphasizing the regulatory role of food on the functions of the body's organs and the body's constitution. Modern nutrition, based on biochemical and physiological research, focuses on the content of macronutrients and micronutrients in food and their impact on specific diseases. Some studies have attempted to combine the two theoretical systems for health management, such as generating dietary plans for the elderly that balance chronic disease control and personal preferences. However, these studies primarily focus on how to apply suggestions from both theories simultaneously, rather than resolving the contradictions between them. When a user has both a clear disease diagnosis and a specific TCM constitution type, a fundamental conflict may arise between the suitability assessment derived from disease-based nutritional management and the suitability assessment derived from TCM constitution conditioning for the same food.

[0004] However, existing technologies are significantly inadequate in handling direct conflicts stemming from different medical theoretical systems. According to authoritative reviews in the field of health informatics, while recommender systems face numerous challenges, including user trust, data privacy, and human-machine collaboration, methods for resolving conflicts in recommendations between Traditional Chinese Medicine (TCM) and modern medical theories have not been adequately explored in the literature. Existing dietary recommendation systems are either built upon a single theoretical framework or simply present recommendations from different sources side-by-side to users, lacking a technological solution capable of identifying, quantifying, and intelligently arbitrating such cross-theoretical conflicts. This technological gap leaves users bewildered when faced with conflicting health advice and limits the in-depth application of integrated medical concepts in digital health management. More importantly, TCM constitution suitability scores and disease-based nutritional suitability scores originate from drastically different epistemological frameworks. The former is based on a holistic, systemic functional model, while the latter is based on empirical, biochemical evidence. Designing a decision-making model to meaningfully weigh and compare these two fundamentally different types of inputs constitutes a challenging technical problem in itself. Summary of the Invention

[0005] To address the aforementioned technical issues, this application provides a dietary recommendation method and system based on multi-dimensional data analysis, which improves the accuracy and rationality of dietary recommendations.

[0006] In a first aspect, embodiments of this application provide a dietary recommendation method based on multi-dimensional data analysis, including: Acquire and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base; Based on the set of TCM theoretical attributes and the user's TCM constitution type, a first rule matching is performed in the preset TCM dietary therapy rule library, and the compatibility score between the target food and the user's TCM constitution is calculated based on the first rule matching result. Based on the set of modern nutritional attributes and the user's list of diagnosed diseases, a second rule matching is performed in a preset disease nutrition rule base, and the disease nutrition suitability score between the target food and the user is calculated based on the second rule matching result. The system determines whether there is a conflict in dietary recommendations based on the TCM constitution suitability score and the disease nutritional suitability score. If there is no conflict in dietary recommendations, a first dietary recommendation result for the target food is generated based on the TCM constitution suitability score and the disease nutritional suitability score. If there is a conflict in dietary recommendations, the severity of the disease and the deviation of the constitution will be assessed based on the user's list of diagnosed diseases and TCM constitution type, respectively, to obtain the corresponding disease severity index and constitution deviation index. Based on the disease severity index and the body constitution deviation index, the TCM body constitution suitability score and the disease nutritional suitability score are weighted and summed, and a second dietary recommendation result for the target food is generated based on the weighted summation result.

[0007] This application provides a dietary recommendation method based on multi-dimensional data analysis. By comprehensively considering both traditional Chinese medicine (TCM) theory and modern nutrition standards, and combining multi-parameter comprehensive decision-making techniques, it achieves truly personalized recommendations, thereby effectively improving the accuracy and practicality of dietary recommendations. In modern society, people's health conditions are increasingly complex, often suffering from certain specific diseases while possessing a particular TCM constitution. In such cases, relying solely on a single theoretical system may not fully meet the needs, and may even produce contradictory recommendation results. This embodiment introduces a TCM constitution suitability score and a disease-nutritional suitability score, which not only objectively reflects the suitability of food in different dimensions but also dynamically detects potential recommendation conflicts, ensuring that the final dietary recommendations are more scientific and reasonable. Furthermore, this embodiment also introduces a disease severity assessment and constitution deviation assessment mechanism, which can flexibly adjust the weight allocation when dietary recommendation conflicts occur, determine the focus of dietary recommendations based on the user's own situation, and resolve conflicts through weighted summation to obtain the final recommendation result. Compared to traditional single-theory recommendation systems, this embodiment significantly enhances the personalization and comprehensiveness of health management, helping users make optimal choices in complex health states, and also provides an important reference for future smart healthcare and big data-driven health interventions.

[0008] Furthermore, the step of acquiring and extracting the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base includes: The system acquires text data or image data and inputs it into a preset food analysis agent, so that the food analysis agent can call the corresponding image recognition algorithm or text parsing algorithm to determine the corresponding target food based on the text data or image data. Based on the target food, extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base; The set of modern nutritional properties includes the energy value per unit mass of food, protein content, fat content, carbohydrate content, dietary fiber content, glycemic index, and the content of various vitamins; The set of TCM theoretical attributes includes the nature, flavor, meridian tropism, and efficacy description of the target food.

[0009] This embodiment further refines the acquisition and analysis process of the target food, utilizing intelligent methods to automatically extract multiple attributes of the food. This approach greatly enhances the practicality and user experience of this embodiment, as users can easily and quickly obtain professional dietary advice by taking photos or entering text, without needing in-depth professional knowledge. Furthermore, through the refined classification and processing of food attributes, this embodiment can better adapt to diverse individual differences, thereby improving the accuracy and rationality of dietary recommendations.

[0010] In one possible implementation, the step of performing a first rule matching in a preset TCM dietary therapy rule base based on the set of TCM theoretical attributes and the user's TCM constitution type, and calculating the compatibility score between the target food and the user's TCM constitution based on the first rule matching result, includes: Based on the set of TCM theoretical attributes and the user's TCM constitution type, a search is performed in the TCM dietary therapy rule base to obtain several TCM constitution suitability rules. Each TCM constitution suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any TCM theoretical attribute and any TCM constitution type, and the consequent conclusions are the suitability level and confidence level of the TCM theoretical attribute and TCM constitution type in the antecedent conditions. Based on the consequent conclusions in each of the aforementioned TCM constitution suitability rules, the TCM constitution suitability score between the target food and the user is calculated.

[0011] This application provides a method for generating a Traditional Chinese Medicine (TCM) constitution suitability score. Based on the TCM theoretical attributes of food and the user's TCM constitution type, rules are matched against a pre-defined TCM dietary therapy rule base. The suitability of the target food with the user's TCM constitution is determined according to the matched rules, and a corresponding score is generated. This embodiment not only fully utilizes the rich resources in the TCM dietary therapy rule base but also introduces the concepts of suitability level and confidence level, making the scoring process more scientifically based and operable. Compared to previous recommendation methods relying on subjective experience, this data- and algorithm-based quantitative analysis method significantly improves the reliability and transparency of the results. Furthermore, due to the adoption of a multi-level antecedent and consequent structure, this embodiment can flexibly handle various complex constitution-food interaction scenarios, avoiding the risk of misjudgment caused by rigid rules and improving the accuracy and rationality of dietary recommendations.

[0012] Furthermore, the step of calculating the TCM constitution compatibility score between the target food and the user based on the consequent conclusions in each of the TCM constitution compatibility rules includes: The fitness level and confidence level in the consequent conclusion of each of the TCM constitution suitability rules are multiplied to obtain the corresponding rule fitness score. The initial TCM constitution suitability score is obtained by summing the suitability scores of each rule. Based on the user's TCM constitution type and the preset constitution-organ correlation matrix, determine the correlation strength between the user's TCM constitution type and each organ. Based on the user's TCM constitution type and the correlation strength between each organ and the food meridian attribute in the TCM theory attribute set, the degree of matching between the target food and the user's organs is determined. The initial TCM constitution compatibility score is corrected based on the degree of matching between the target food and the user's internal organs, and the corrected score is normalized to a preset numerical range to obtain the TCM constitution compatibility score.

[0013] This application provides a method for calculating a Traditional Chinese Medicine (TCM) constitution suitability score. Based on an initial TCM constitution suitability score obtained according to TCM constitution suitability rules, it introduces a constitution-organ correlation matrix and food meridian attributes to further refine the initial TCM constitution suitability score, making the score more closely reflect the individual user's actual situation. This application considers that different TCM constitution types have different correlation strengths with various organs, and that the meridian attributes of foods indicate their impact on specific organs. Therefore, the influence of food meridian attributes is amplified or diminished depending on the user's TCM constitution type. With this in mind, this application refines the initial TCM constitution suitability score by determining the degree of matching between target foods and the user's organs, more accurately predicting the comprehensive impact of food on the human body, and thus providing more targeted dietary recommendations, improving the accuracy and rationality of dietary recommendations.

[0014] In one possible implementation, the step of performing a second rule matching in a preset disease nutrition rule base based on the set of modern nutritional attributes and the user's list of diagnosed diseases, and calculating the disease nutrition suitability score between the target food and the user based on the second rule matching result, includes: Iterate through each confirmed disease in the list of confirmed diseases. For any confirmed disease, search the disease nutrition rule base according to the confirmed disease and match the corresponding disease nutrition suitability rule. The disease nutrition suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any disease name. The consequent conclusions include each recommended nutrient, the intake range of each recommended nutrient, each restricted nutrient, and the restricted intake range of each restricted nutrient. The set of modern nutritional attributes is compared with the nutritional suitability rules for each of the diseases, and the suitability score between the target food and each of the diagnosed diseases is determined based on the comparison results. The disease nutritional suitability score between the target food and the user is calculated based on each of the aforementioned suitability scores.

[0015] This application provides a method for generating a disease nutritional suitability score. By systematically reviewing diagnosed diseases and matching them with a disease nutrition rule base, a highly personalized nutritional intervention strategy is achieved. This embodiment first matches each disease in the disease nutrition rule base to obtain the corresponding disease nutritional suitability rule. Then, it compares the set of modern nutritional attributes of food with each disease nutritional suitability rule. For example, it compares whether the food contains recommended nutrients, whether it contains restricted nutrients, whether the nutrient content is within the recommended intake range, and whether it exceeds the restricted intake range. This approach fully considers the multiple pathological characteristics of patients with chronic diseases and the diverse nutritional properties of food itself. It avoids the excessive focus on a single pathological factor in traditional methods and instead adopts a systemic approach to address complex health issues, improving the accuracy and rationality of dietary recommendations.

[0016] In one possible implementation, the step of determining whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutritional suitability score, and if no dietary recommendation conflict exists, generating a first dietary recommendation result for the target food based on the TCM constitution suitability score and the disease nutritional suitability score, includes: The dietary recommendation score is obtained by multiplying the TCM constitution suitability score and the disease nutrition suitability score. If the dietary recommendation score is less than 0 and the absolute values ​​of the TCM constitution suitability score and the disease nutrition suitability score are both greater than a preset threshold, then a dietary recommendation conflict is determined to exist; otherwise, no dietary recommendation conflict is determined to exist. When no dietary recommendation conflict is determined, if both the TCM constitution suitability score and the disease nutritional suitability score are positive, a positive dietary recommendation result for the target food is generated; if both the TCM constitution suitability score and the disease nutritional suitability score are negative, a negative dietary recommendation result for the target food is generated; if the signs of the TCM constitution suitability score and the disease nutritional suitability score are different, the generation of a positive or negative dietary recommendation result is determined based on the absolute values ​​of the TCM constitution suitability score and the disease nutritional suitability score.

[0017] This application provides an efficient dietary recommendation conflict detection mechanism. It automates the identification of recommendation conflicts through product operations and threshold settings, avoiding contradictory food recommendation results and improving the user experience. Furthermore, this embodiment also provides a clear decision-making path in non-conflict situations. It determines whether to generate a positive or negative dietary recommendation result based on the sign and absolute value of the Traditional Chinese Medicine constitution suitability score and the disease nutritional suitability score, thereby improving the accuracy and rationality of dietary recommendations.

[0018] In one possible implementation, if a dietary recommendation conflict exists, the user's disease severity and constitution deviation are assessed based on their diagnosed disease list and TCM constitution type, respectively, to obtain corresponding disease severity and constitution deviation indicators, including: Based on the list of confirmed diseases and the preset disease risk classification table, determine the basic risk score for each confirmed disease in the list of confirmed diseases; Based on the user's physiological index data, the actual control status of each of the diagnosed diseases is evaluated, and the control status correction coefficient of each diagnosed disease is obtained. The severity index of the disease is calculated based on the basic risk score and control status correction coefficient of each diagnosed disease. The user's basic constitution deviation index is determined based on the TCM constitution type and the preset constitution deviation mapping table; Obtain the TCM constitution assessment form uploaded by the user, and determine the user's current constitution characteristic score based on the TCM constitution assessment form; The physical fitness deviation index is calculated based on the basic physical fitness deviation index and the current physical fitness characteristic score.

[0019] This application provides a method for generating disease severity indicators and constitution deviation indicators. On one hand, by combining a static risk grading table, a list of diagnosed diseases, and dynamic physiological indicator data, the urgency of a user's current disease is determined, and the corresponding disease severity indicator is calculated. On the other hand, the introduction of constitution deviation indicators makes the quantitative expression of traditional Chinese medicine theory a reality. Through static TCM constitution types, constitution deviation mapping tables, and dynamic TCM constitution scales, the urgency of a user's current constitution deviation is determined, and the corresponding constitution deviation indicator is calculated. The real-time calculation of disease severity indicators and constitution deviation indicators allows this embodiment to update the user's health status profile in real time based on these indicators, ensuring the timeliness and adaptability of the recommended plan and improving the accuracy and rationality of dietary recommendations.

[0020] Furthermore, the dietary recommendation method also includes generating corresponding recommendation text based on the second dietary recommendation result when a second dietary recommendation result for the target food is generated based on the weighted summation result, and sending the second dietary recommendation result and the recommendation text to a designated device, specifically: Based on the TCM constitution suitability score and the user's TCM constitution type, generate the first evaluation text of the target food in the TCM dimension; Based on the disease nutritional suitability score and the user's list of diagnosed diseases, a second evaluation text of the target food in terms of nutrition is generated. Based on the disease severity index, the body constitution deviation index, and the second dietary recommendation result, an explanatory text for the dietary recommendation conflict is generated; The recommended text is generated by combining the first evaluation text, the second evaluation text, and the explanation text, and the second dietary recommendation result and the recommended text are sent to the designated device.

[0021] This application provides a method for generating recommendation text, transforming abstract ratings into concrete textual descriptions to generate detailed recommendation text for user reference. By generating a first evaluation text, a second evaluation text, and an explanatory text, and combining these texts to generate the recommendation text, the method clearly demonstrates the food recommendation logic and conflict resolution logic of this embodiment to the user. This helps to enhance the user's understanding and trust in the recommendation logic, reduce doubts or resistance caused by information asymmetry, and improve the user experience.

[0022] In one possible implementation, the step of weighted summing of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generating a second dietary recommendation result for the target food based on the weighted summation result, includes: Based on the disease severity index and the constitution deviation index, the TCM constitution suitability score and the disease nutritional suitability score are weighted and summed to obtain a comprehensive score. An initial second dietary recommendation for the target food is generated based on the numerical range to which the comprehensive score belongs; If the disease status field in the confirmed disease list and the user's user information meet the preset conditions, then the preset priority rule or correction coefficient is invoked to correct the initial second dietary recommendation result, and the second dietary recommendation result is generated; otherwise, the initial second dietary recommendation result is used as the second dietary recommendation result.

[0023] Secondly, embodiments of this application provide a diet recommendation system based on multi-dimensional data analysis, including an acquisition module, a first matching module, a second matching module, a conflict judgment module, an evaluation module, and a conflict resolution module; The acquisition module is used to acquire and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base, based on the target food. The first matching module is used to perform a first rule matching in a preset TCM dietary therapy rule library based on the set of TCM theoretical attributes and the user's TCM constitution type, and calculate the compatibility score between the target food and the user's TCM constitution based on the first rule matching result; The second matching module is used to perform second rule matching in a preset disease nutrition rule base according to the set of modern nutritional attributes and the user's list of diagnosed diseases, and calculate the disease nutrition suitability score between the target food and the user based on the second rule matching result; The conflict judgment module is used to determine whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutrition suitability score. If there is no dietary recommendation conflict, a first dietary recommendation result for the target food is generated based on the TCM constitution suitability score and the disease nutrition suitability score. The assessment module is used to assess the severity of the disease and the deviation of the constitution of the user based on the user's list of diagnosed diseases and TCM constitution type if there is a conflict in dietary recommendations, and to obtain the corresponding disease severity index and constitution deviation index. The conflict resolution module is used to perform a weighted summation of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generate a second dietary recommendation result for the target food based on the weighted summation result. Attached Figure Description

[0024] Figure 1 A flowchart illustrating a dietary recommendation method based on multi-dimensional data analysis provided in this application embodiment; Figure 2 This is a flowchart illustrating the process of obtaining multi-dimensional data in a dietary recommendation method based on multi-dimensional data analysis, as provided in an embodiment of this application. Figure 3 A flowchart illustrating the calculation of a TCM constitution suitability score in a dietary recommendation method based on multi-dimensional data analysis, provided in an embodiment of this application; Figure 4 A flowchart illustrating the calculation of disease nutritional suitability score in a dietary recommendation method based on multi-dimensional data analysis, provided in an embodiment of this application; Figure 5A schematic diagram illustrating the conflict identification process in a diet recommendation method based on multi-dimensional data analysis provided in this application embodiment; Figure 6 A flowchart illustrating the weighted summation of TCM constitution suitability score and disease nutrition suitability score in a dietary recommendation method based on multi-dimensional data analysis provided in this application embodiment; Figure 7 A flowchart illustrating the generation of a second recommendation result and recommendation text in a diet recommendation method based on multi-dimensional data analysis provided in this application embodiment; Figure 8 This is a flowchart illustrating the handling of special scenarios in a dietary recommendation method based on multi-dimensional data analysis provided in an embodiment of this application. Figure 9 This is a schematic diagram of the structure of a diet recommendation system based on multi-dimensional data analysis, provided in an embodiment of this application. Detailed Implementation

[0025] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0026] It should be noted that the step numbers in this document are only for the convenience of explaining the specific embodiments and are not intended to limit the order in which the steps are performed. In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of that feature.

[0027] Example 1: like Figure 1 As shown, Example 1 provides a dietary recommendation method based on multi-dimensional data analysis, including steps S1-S6: Step S1: Obtain and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from the preset food knowledge base; Step S2: Based on the set of TCM theoretical attributes and the user's TCM constitution type, perform a first rule matching in the preset TCM dietary therapy rule library, and calculate the compatibility score between the target food and the user's TCM constitution based on the first rule matching result; Step S3: Based on the set of modern nutritional attributes and the user's list of diagnosed diseases, perform a second rule matching in the preset disease nutrition rule base, and calculate the disease nutrition suitability score between the target food and the user based on the second rule matching result; Step S4: Determine whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutritional suitability score. If there is no dietary recommendation conflict, generate a first dietary recommendation result for the target food based on the TCM constitution suitability score and the disease nutritional suitability score. Step S5: If there is a conflict in dietary recommendations, the severity of the disease and the deviation of the constitution will be assessed based on the user's list of diagnosed diseases and TCM constitution type, respectively, to obtain the corresponding disease severity index and constitution deviation index. Step S6: Based on the disease severity index and the constitution deviation index, perform a weighted summation of the TCM constitution suitability score and the disease nutritional suitability score, and generate a second dietary recommendation result for the target food based on the weighted summation result.

[0028] This application provides a dietary recommendation method based on multi-dimensional data analysis. By comprehensively considering both traditional Chinese medicine (TCM) theory and modern nutrition standards, and combining multi-parameter comprehensive decision-making techniques, it achieves truly personalized recommendations, thereby effectively improving the accuracy and practicality of dietary recommendations. In modern society, people's health conditions are increasingly complex, often suffering from certain specific diseases while possessing a particular TCM constitution. In such cases, relying solely on a single theoretical system may not fully meet the needs, and may even produce contradictory recommendation results. This embodiment introduces a TCM constitution suitability score and a disease-nutritional suitability score, which not only objectively reflects the suitability of food in different dimensions but also dynamically detects potential recommendation conflicts, ensuring that the final dietary recommendations are more scientific and reasonable. Furthermore, this embodiment also introduces a disease severity assessment and constitution deviation assessment mechanism, which can flexibly adjust the weight allocation when dietary recommendation conflicts occur, determine the focus of dietary recommendations based on the user's own situation, and resolve conflicts through weighted summation to obtain the final recommendation result. Compared to traditional single-theory recommendation systems, this embodiment significantly enhances the personalization and comprehensiveness of health management, helping users make optimal choices in complex health states, and also provides an important reference for future smart healthcare and big data-driven health interventions.

[0029] Furthermore, in step S1, the step of acquiring and extracting the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base includes: The system acquires text data or image data and inputs it into a preset food analysis agent, so that the food analysis agent can call the corresponding image recognition algorithm or text parsing algorithm to determine the corresponding target food based on the text data or image data. Based on the target food, extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base; The set of modern nutritional properties includes the energy value per unit mass of food, protein content, fat content, carbohydrate content, dietary fiber content, glycemic index, and the content of various vitamins; The set of TCM theoretical attributes includes the nature, flavor, meridian tropism, and efficacy description of the target food.

[0030] This embodiment further refines the acquisition and analysis process of the target food, utilizing intelligent methods to automatically extract multiple attributes of the food. This approach greatly enhances the practicality and user experience of this embodiment, as users can easily and quickly obtain professional dietary advice by taking photos or entering text, without needing in-depth professional knowledge. Furthermore, through the refined classification and processing of food attributes, this embodiment can better adapt to diverse individual differences, thereby improving the accuracy and rationality of dietary recommendations.

[0031] In a preferred embodiment, such as Figure 2 As shown, this embodiment first extracts the user's health-related information from the user profile database through a data interface. This information includes the user's TCM constitution type, list of diagnosed diseases, allergy history records, and current physiological indicator data. During the data acquisition phase, this embodiment encodes the user's TCM constitution type into standardized constitution codes, such as encoding "Yang Deficiency Constitution" as "YangX", "Yin Deficiency Constitution" as "YinX", and "Phlegm-Dampness Constitution" as "TanS", totaling nine basic constitution types. For disease information, it is encoded and stored according to the International Classification of Diseases 10th Revision (ICD-10), with each disease record including fields such as disease code, disease name, diagnosis time, current control status, and severity level. Simultaneously, key physiological indicators are extracted from the user's most recent physical examination record, including but not limited to fasting blood glucose, glycated hemoglobin level, blood pressure, blood lipid indicators, and liver and kidney function indicators. These numerical data are stored in standard units and accompanied by a collection timestamp.

[0032] When acquiring target food information, this embodiment receives a food identifier from the food analysis agent. This identifier corresponds to a unique food entry in the system's knowledge base. Each entry contains two parallel attribute sets: modern nutritional attributes and traditional Chinese medicine (TCM) attributes. Modern nutritional attributes are represented by structured data, including quantitative indicators such as energy value per 100 grams (kcal), protein content (grams), fat content (grams), carbohydrate content (grams), dietary fiber content (grams), major vitamin content, major mineral content, and glycemic index (GI). TCM attributes are represented by labels, including qualitative information such as the food's nature (five categories: cold, cool, neutral, warm, and hot), taste (five flavors: sour, bitter, sweet, pungent, and salty), meridian tropism (twelve meridians including liver, heart, spleen, lung, and kidney), and efficacy descriptions. This embodiment loads these heterogeneous data into a unified in-memory data structure to prepare data for subsequent scoring calculations. The food analysis agent is a functional module specifically designed to identify and analyze food information. This agent can acquire basic food information through image recognition, text parsing, or user input, and assign a unique identifier to each food item. As the front-end data acquisition module in this embodiment, the food analysis agent is responsible for converting the food consumed or planned by the user into structured data recognizable by this embodiment.

[0033] In one possible implementation, in step S2, the first rule matching is performed in a preset TCM dietary therapy rule base based on the TCM theoretical attribute set and the user's TCM constitution type, and the compatibility score between the target food and the user's TCM constitution is calculated based on the first rule matching result, including: Based on the set of TCM theoretical attributes and the user's TCM constitution type, a search is performed in the TCM dietary therapy rule base to obtain several TCM constitution suitability rules. Each TCM constitution suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any TCM theoretical attribute and any TCM constitution type, and the consequent conclusions are the suitability level and confidence level of the TCM theoretical attribute and TCM constitution type in the antecedent conditions. Based on the consequent conclusions in each of the aforementioned TCM constitution suitability rules, the TCM constitution suitability score between the target food and the user is calculated.

[0034] This application provides a method for generating a Traditional Chinese Medicine (TCM) constitution suitability score. Based on the TCM theoretical attributes of food and the user's TCM constitution type, rules are matched against a pre-defined TCM dietary therapy rule base. The suitability of the target food with the user's TCM constitution is determined according to the matched rules, and a corresponding score is generated. This embodiment not only fully utilizes the rich resources in the TCM dietary therapy rule base but also introduces the concepts of suitability level and confidence level, making the scoring process more scientifically based and operable. Compared to previous recommendation methods relying on subjective experience, this data- and algorithm-based quantitative analysis method significantly improves the reliability and transparency of the results. Furthermore, due to the adoption of a multi-level antecedent and consequent structure, this embodiment can flexibly handle various complex constitution-food interaction scenarios, avoiding the risk of misjudgment caused by rigid rules and improving the accuracy and rationality of dietary recommendations.

[0035] Furthermore, the step of calculating the TCM constitution compatibility score between the target food and the user based on the consequent conclusions in each of the TCM constitution compatibility rules includes: The fitness level and confidence level in the consequent conclusion of each of the TCM constitution suitability rules are multiplied to obtain the corresponding rule fitness score. The initial TCM constitution suitability score is obtained by summing the suitability scores of each rule. Based on the user's TCM constitution type and the preset constitution-organ correlation matrix, determine the correlation strength between the user's TCM constitution type and each organ. Based on the user's TCM constitution type and the correlation strength between each organ and the food meridian attribute in the TCM theory attribute set, the degree of matching between the target food and the user's organs is determined. The initial TCM constitution compatibility score is corrected based on the degree of matching between the target food and the user's internal organs, and the corrected score is normalized to a preset numerical range to obtain the TCM constitution compatibility score.

[0036] This application provides a method for calculating a Traditional Chinese Medicine (TCM) constitution suitability score. Based on an initial TCM constitution suitability score obtained according to TCM constitution suitability rules, it introduces a constitution-organ correlation matrix and food meridian attributes to further refine the initial TCM constitution suitability score, making the score more closely reflect the individual user's actual situation. This application considers that different TCM constitution types have different correlation strengths with various organs, and that the meridian attributes of foods indicate their impact on specific organs. Therefore, the influence of food meridian attributes is amplified or diminished depending on the user's TCM constitution type. With this in mind, this application refines the initial TCM constitution suitability score by determining the degree of matching between target foods and the user's organs, more accurately predicting the comprehensive impact of food on the human body, and thus providing more targeted dietary recommendations, improving the accuracy and rationality of dietary recommendations.

[0037] In a preferred embodiment, such as Figure 3 As shown, this embodiment calculates a TCM constitution compatibility score S based on the user's TCM constitution type and the TCM attributes of the target food. tcm The scoring method combines knowledge base matching and rule-based reasoning. This embodiment first retrieves a set of dietary guidelines related to the user's constitution type from a traditional Chinese medicine dietary therapy rule base. This rule base is stored in an "IF-THEN" structure, where each rule includes an antecedent condition (a combination of the user's constitution type and food attributes) and a consequent conclusion (suitability evaluation and its confidence level). For example, the rule "IF user constitution = Yang deficiency AND food nature = warm THEN suitability = high CONFIDENCE = 0.9" indicates that warm foods have a high suitability for users with a Yang deficiency constitution, with a confidence level of 0.9.

[0038] By traversing the rule base, all rules whose antecedent conditions match the current user's constitution and the target food attributes are extracted to form a candidate rule set. For each matching rule, a base score is assigned based on the suitability evaluation (divided into four levels: high, medium, low, and contraindicated) in its consequent conclusion. Specifically, a rule with "high" suitability contributes a positive score of +2.0, a rule with "medium" suitability contributes a positive score of +1.0, a rule with "low" suitability contributes a negative score of -1.0, and a rule with "contraindicated" suitability contributes a negative score of -3.0. Each base score is multiplied by the rule's confidence coefficient to obtain the rule's weighted contribution value. Finally, the weighted contribution values ​​of all matching rules are summed and divided by the total number of matching rules to obtain a preliminary TCM constitution suitability score.

[0039] To enhance the accuracy of the scoring, this embodiment further incorporates a correction based on the matching degree between food meridian tropism and the user's constitution and susceptible organs. Traditional Chinese medicine (TCM) theory posits that different constitution types exhibit specific organ weakness tendencies; for example, Yang deficiency often involves spleen and kidney Yang deficiency, while Yin deficiency often involves liver and kidney Yin deficiency. This embodiment maintains a constitution-organ correlation matrix, which is a pre-constructed knowledge base based on classic TCM theories such as the *Huangdi Neijing* and *Classification and Determination of TCM Constitutions*. The matrix construction process includes: First, a team of TCM experts, based on TCM constitution theory, identifies the physiological and pathological correlations between nine constitution types and the five internal organs (heart, liver, spleen, lungs, and kidneys) and the six viscera (gallbladder, stomach, small intestine, large intestine, bladder, and triple burner); second, the Delphi method is used to conduct multiple rounds of consultation and quantification of expert opinions, transforming qualitative TCM theories into correlation strength values ​​between 0 and 1, where 0 represents no correlation and 1 represents a strong correlation; finally, a 9×11 dimensional correlation matrix is ​​formed, where each element represents the degree of influence of a specific constitution type on a specific organ. For example, the correlation strength between Yang deficiency constitution and the kidneys is 0.9, and the correlation strength with the spleen is 0.7, reflecting the traditional Chinese medicine understanding that Yang deficiency constitution mainly involves insufficient kidney Yang, while also affecting spleen Yang. This matrix serves as the system's basic knowledge base, loaded during system initialization, and can be periodically optimized and updated based on clinical practice data. When the meridian tropism of the target food highly matches the susceptible organs of the user's constitution, the system applies a positive correction factor to the initial score, with a correction range between 0.2 and 0.5, the specific value depending on the correlation strength coefficient. The final TCM constitution suitability score S... tcm Normalized to the range [-5, +5], positive values ​​indicate that the food is suitable for the user's constitution, negative values ​​indicate that the food is not suitable for the user's constitution, and the absolute value indicates the strength of the compatibility.

[0040] In one possible implementation, in step S3, the second rule matching is performed in a preset disease nutrition rule base based on the set of modern nutritional attributes and the user's list of diagnosed diseases. The result of the second rule matching is used to calculate the disease nutrition suitability score between the target food and the user, including: Iterate through each confirmed disease in the list of confirmed diseases. For any confirmed disease, search the disease nutrition rule base according to the confirmed disease and match the corresponding disease nutrition suitability rule. The disease nutrition suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any disease name. The consequent conclusions include each recommended nutrient, the intake range of each recommended nutrient, each restricted nutrient, and the restricted intake range of each restricted nutrient. The set of modern nutritional attributes is compared with the nutritional suitability rules for each of the diseases, and the suitability score between the target food and each of the diagnosed diseases is determined based on the comparison results. The disease nutritional suitability score between the target food and the user is calculated based on each of the aforementioned suitability scores.

[0041] This application provides a method for generating a disease nutritional suitability score. By systematically reviewing diagnosed diseases and matching them with a disease nutrition rule base, a highly personalized nutritional intervention strategy is achieved. This embodiment first matches each disease in the disease nutrition rule base to obtain the corresponding disease nutritional suitability rule. Then, it compares the set of modern nutritional attributes of food with each disease nutritional suitability rule. For example, it compares whether the food contains recommended nutrients, whether it contains restricted nutrients, whether the nutrient content is within the recommended intake range, and whether it exceeds the restricted intake range. This approach fully considers the multiple pathological characteristics of patients with chronic diseases and the diverse nutritional properties of food itself. It avoids the excessive focus on a single pathological factor in traditional methods and instead adopts a systemic approach to address complex health issues, improving the accuracy and rationality of dietary recommendations.

[0042] In a preferred embodiment, such as Figure 4 As shown, this embodiment calculates the compatibility score between the target food and the diagnosed disease for each user's confirmed disease, and then calculates a comprehensive disease nutritional compatibility score S based on each compatibility score. disease .

[0043] For calculating the suitability score for a single disease, the system first retrieves the corresponding disease nutrition suitability rules from the disease nutrition management knowledge base. This knowledge base is built based on evidence-based medicine and clearly indicates the recommended intake range, restricted intake components, and prohibited food lists for patients with specific diseases. Then, the nutritional data of the target food is compared item by item with the disease nutrition suitability rules. For nutrients that need to be restricted, the system calculates the ratio of the actual content of that component in the food to the recommended limit. If the ratio is less than 0.5, it indicates that the food is beneficial for disease management in this nutritional dimension, contributing a positive score. If the ratio is between 0.5 and 1.0, it indicates that the food can be consumed in moderation, contributing a neutral score. If the ratio is greater than 1.0, it indicates that the food may be detrimental to disease control, contributing a negative score, with a higher ratio resulting in a higher negative score. For example, for diabetic patients, the system focuses on assessing the glycemic index and carbohydrate content of foods. If the glycemic index of a food is below 55 (low GI food), the system assigns a positive score of +1.5. Foods with a glycemic index (GI) between 55 and 70 (medium GI) are assigned a score of 0. Foods with a GI above 70 (high GI) are assigned a negative score of -2.0. For nutrients requiring increased intake, the opposite scoring logic applies. Foods rich in the nutrient (at levels exceeding 20% ​​of the recommended daily intake) are assigned a positive score. Finally, the scores for all nutritional dimensions are weighted and summed, with weight coefficients reflecting the relative importance of each nutrient for disease management. The total weights are normalized to 1.0. The weighted sum yields the food's suitability score for a specific disease.

[0044] Finally, considering that users may suffer from multiple diseases simultaneously, it is necessary to aggregate the individual scores of each disease into a comprehensive disease-nutritional suitability score. The aggregation method can be set according to actual needs, such as direct addition or weighted summation based on the severity of the disease. This comprehensive score is also normalized to the range of [-5, +5], with positive values ​​indicating that the food is beneficial to the user's disease management and negative values ​​indicating that the food may be detrimental to disease control.

[0045] In one possible implementation, in step S4, determining whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutritional suitability score, and if no dietary recommendation conflict exists, generating a first dietary recommendation result for the target food based on the TCM constitution suitability score and the disease nutritional suitability score, includes: The dietary recommendation score is obtained by multiplying the TCM constitution suitability score and the disease nutrition suitability score. If the dietary recommendation score is less than 0 and the absolute values ​​of the TCM constitution suitability score and the disease nutrition suitability score are both greater than a preset threshold, then a dietary recommendation conflict is determined to exist; otherwise, no dietary recommendation conflict is determined to exist. When no dietary recommendation conflict is determined, if both the TCM constitution suitability score and the disease nutritional suitability score are positive, a positive dietary recommendation result for the target food is generated; if both the TCM constitution suitability score and the disease nutritional suitability score are negative, a negative dietary recommendation result for the target food is generated; if the signs of the TCM constitution suitability score and the disease nutritional suitability score are different, the generation of a positive or negative dietary recommendation result is determined based on the absolute values ​​of the TCM constitution suitability score and the disease nutritional suitability score.

[0046] This application provides an efficient dietary recommendation conflict detection mechanism. It automates the identification of recommendation conflicts through product operations and threshold settings, avoiding contradictory food recommendation results and improving the user experience. Furthermore, this embodiment also provides a clear decision-making path in non-conflict situations. It determines whether to generate a positive or negative dietary recommendation result based on the sign and absolute value of the Traditional Chinese Medicine constitution suitability score and the disease nutritional suitability score, thereby improving the accuracy and rationality of dietary recommendations.

[0047] In a preferred embodiment, such as Figure 5 As shown, in obtaining the TCM constitution suitability score Nutritional suitability score for diseases Next, a conflict detection algorithm is executed to determine whether the two ratings constitute a recommendation conflict. The core criteria for conflict detection are the opposite signs of the ratings and the absolute value threshold. Specifically, it first checks whether the signs of the two ratings are opposite, i.e., whether the following conditions are met. The condition is that if this condition is met, it indicates that the evaluation of food on the two health dimensions is in opposite directions, and there is a potential conflict.

[0048] Simply having opposite signs is insufficient to confirm a conflict, because a smaller absolute value of the score indicates a weaker recommendation strength for that dimension, and such disagreement is insufficient to constitute a substantial conflict requiring arbitration. Therefore, it is necessary to further examine whether the absolute values ​​of both scores exceed a preset significance threshold. This threshold is determined through large-scale historical data statistical analysis; the default threshold set in this embodiment is 1.5. Only when... and Only when the two scores are considered sufficiently significant can they be considered to be meaningful.

[0049] Based on the above two conditions, the conflict determination logic can be expressed as follows: If a recommendation conflict is detected, a conflict arbitration process is initiated. If no conflict is detected (ELSE), the standard recommendation result is output directly. In the case of no conflict, if both ratings are positive, the food is marked as "recommended." If both ratings are negative, the food is marked as "avoid." If the two ratings have opposite signs, the recommendation decision is made based on the rating with the larger absolute value.

[0050] In one possible implementation, in step S5, if a dietary recommendation conflict exists, the user's disease severity and constitution deviation are assessed based on their diagnosed disease list and TCM constitution type, respectively, to obtain corresponding disease severity indicators and constitution deviation indicators, including: Based on the list of confirmed diseases and the preset disease risk classification table, determine the basic risk score for each confirmed disease in the list of confirmed diseases; Based on the user's physiological index data, the actual control status of each of the diagnosed diseases is evaluated, and the control status correction coefficient of each diagnosed disease is obtained. The severity index of the disease is calculated based on the basic risk score and control status correction coefficient of each diagnosed disease. The user's basic constitution deviation index is determined based on the TCM constitution type and the preset constitution deviation mapping table; Obtain the TCM constitution assessment form uploaded by the user, and determine the user's current constitution characteristic score based on the TCM constitution assessment form; The physical fitness deviation index is calculated based on the basic physical fitness deviation index and the current physical fitness characteristic score.

[0051] This application provides a method for generating disease severity indicators and constitution deviation indicators. On one hand, by combining a static risk grading table, a list of diagnosed diseases, and dynamic physiological indicator data, the urgency of a user's current disease is determined, and the corresponding disease severity indicator is calculated. On the other hand, the introduction of constitution deviation indicators makes the quantitative expression of traditional Chinese medicine theory a reality. Through static TCM constitution types, constitution deviation mapping tables, and dynamic TCM constitution scales, the urgency of a user's current constitution deviation is determined, and the corresponding constitution deviation indicator is calculated. The real-time calculation of disease severity indicators and constitution deviation indicators allows this embodiment to update the user's health status profile in real time based on these indicators, ensuring the timeliness and adaptability of the recommended plan and improving the accuracy and rationality of dietary recommendations.

[0052] In a preferred embodiment, upon determining a recommendation conflict, the first step in initiating the intelligent arbitration process is to calculate two key parameters for dynamic weight adjustment: a disease severity quantification index and a physical fitness deviation index. The disease severity quantification index reflects the degree of threat posed to the user's current disease state to their health; its calculation comprehensively considers both the intrinsic severity of the disease and the user's individual control status.

[0053] To assess the intrinsic severity of a disease, this embodiment maintains a disease risk grading table based on medical consensus. This table categorizes common diseases into five levels according to their threat to life and health, assigning a baseline risk score of 1 to 5. For example, well-controlled hypertension is assigned a score of 2, diabetes a score of 3, coronary heart disease a score of 4, and malignant tumors a score of 5. The baseline risk score is obtained by querying this grading table based on the user's ICD-10 disease code. Then, based on the user's physiological indicator data, the actual control status of the disease is assessed, and the control status correction coefficient is calculated. This correction factor is calculated based on the degree of deviation of key physiological indicators from the target control range. Taking diabetes as an example, the user's fasting plasma glucose (FPG) and glycated hemoglobin (HbA1c) values ​​are extracted and compared with the target control range for diabetes management. If FPG is within the range of 5.0-7.0 mmol / L and HbA1c is below 7.0%, it indicates good control, and the correction factor is [not specified]. The value is 0.8. If FPG is between 7.0-10.0 mmol / L or HbA1c is between 7.0-9.0%, it indicates fair control, and the correction factor is 1.0. If FPG exceeds 10.0 mmol / L or HbA1c exceeds 9.0%, it indicates poor control, and the correction factor is 1.3. Quantitative indicators of disease severity. Calculated as The final value range is 1.0 to 6.5.

[0054] The constitution deviation index is used to quantify the degree of deviation between a user's current constitution and the ideal balanced constitution. Traditional Chinese medicine theory considers the balanced constitution to be the healthiest constitution, while the other eight imbalanced constitutions represent different degrees of health imbalance. This embodiment maintains a constitution deviation mapping table, assigning a basic deviation value to each constitution type. The deviation for the balanced constitution is 0; the deviation for mild imbalanced constitutions such as Qi deficiency, Yang deficiency, and Yin deficiency is 2; the deviation for moderate imbalanced constitutions such as phlegm-dampness and damp-heat is 3; the deviation for severe imbalanced constitutions such as blood stasis and Qi stagnation is 4; and the deviation for the special constitution (allergic constitution) is 5.

[0055] Then, the baseline deviation is further corrected by combining the user's constitution assessment score. During the constitution assessment process, users typically need to complete a standardized Traditional Chinese Medicine (TCM) constitution scale. In this embodiment, the scale score determines the user's main constitution type and the typicality of that constitution. If the user's constitution assessment score shows that their constitution characteristics are very typical (score above the threshold of 80%), an enhancement coefficient of 1.2 times is applied to the baseline deviation. If the constitution characteristics are not very typical (score between 60% and 80%), the baseline deviation remains unchanged. Constitution Deviation Index The final value range is from 0 to 6.0. A higher value indicates a more severe imbalance in the user's constitution, requiring more dietary adjustments to improve their condition. Furthermore, if a user's scores for all constitution types are below 60%, it indicates a balanced constitution or indistinct constitution characteristics, in which case a general health recommendation mode is adopted. In this mode, this embodiment does not rely on the suitability score for a specific constitution, but rather makes recommendation decisions based on the user's basic health indicators. Specifically, it prioritizes the nutritional balance and general health attributes of food, including calorie suitability, the rationality of the three major nutrients ratio, and the richness of vitamins and minerals. Simultaneously, it reduces the weight coefficient of the TCM constitution dimension in multi-objective optimization, adjusting the weight from the default 0.3 to 0.1, and correspondingly increases the weight of nutritional and disease management objectives to ensure that the recommendation results still meet the user's overall health needs. In this case, this embodiment will explain to the user in the recommendation results that the general health recommendation mode is currently being used and suggest that the user undergo a reassessment of their constitution when their constitution characteristics become more pronounced to obtain more personalized dietary recommendations.

[0056] In a preferred embodiment, in step S6, as Figure 6 As shown, after obtaining disease severity indicators Deviation from physical fitness index Then, a dynamic weight calculation algorithm is executed to determine the weight coefficients for the Traditional Chinese Medicine dimension. And nutritional dimension weighting coefficient The core idea of ​​dynamic weight calculation is to find the optimal balance between the safety requirements of disease management and the long-term benefits of physical conditioning based on the user's individual health condition. The weight calculation uses a normalization algorithm based on relative importance ratios. First, the importance factors of the nutritional dimension are calculated. This factor is directly proportional to the disease severity index, and the calculation formula is: ,in The current system is set to use the baseline coefficient based on statistical analysis of a large number of clinical cases. Similarly, the importance factors for the Traditional Chinese Medicine dimension were calculated. ,in As the baseline coefficient, currently set The two importance factors reflect the priority and intensity of the two health dimensions under the current user situation.

[0057] Then, the importance factors are converted into weight coefficients through normalization. The weight coefficients for the nutrition dimension are calculated as follows: The weighting coefficient for the Traditional Chinese Medicine dimension is calculated as follows: Through this normalization process, the sum of the two weighting coefficients is always equal to 1.0, satisfying the mathematical requirements of a weighted average. When a user's disease severity index is high, The corresponding value increases, placing greater emphasis on disease-specific nutritional management when resolving conflicts. When a user's physical condition deviates significantly, The value increases accordingly, indicating a greater emphasis on the need for TCM constitution conditioning when resolving conflicts.

[0058] Finally, the calculated dynamic weighting coefficients are used to perform a weighted average of the two original fit scores to generate a comprehensive score. The calculation formula is: .because and In cases of conflict, the signs are reversed. The weighted average is essentially a trade-off; its value and sign depend on which dimension's score dominates under the current weighting. (Comprehensive Score) The value range is still within the interval [-5, +5], and its sign and absolute value reflect the final recommendation tendency after intelligent arbitration.

[0059] Furthermore, the dietary recommendation method also includes generating corresponding recommendation text based on the second dietary recommendation result when a second dietary recommendation result for the target food is generated based on the weighted summation result, and sending the second dietary recommendation result and the recommendation text to a designated device, specifically: Based on the TCM constitution suitability score and the user's TCM constitution type, generate the first evaluation text of the target food in the TCM dimension; Based on the disease nutritional suitability score and the user's list of diagnosed diseases, a second evaluation text of the target food in terms of nutrition is generated. Based on the disease severity index, the body constitution deviation index, and the second dietary recommendation result, an explanatory text for the dietary recommendation conflict is generated; The recommended text is generated by combining the first evaluation text, the second evaluation text, and the explanation text, and the second dietary recommendation result and the recommended text are sent to the designated device.

[0060] This application provides a method for generating recommendation text, transforming abstract ratings into concrete textual descriptions to generate detailed recommendation text for user reference. By generating a first evaluation text, a second evaluation text, and an explanatory text, and combining these texts to generate the recommendation text, the method clearly demonstrates the food recommendation logic and conflict resolution logic of this embodiment to the user. This helps to enhance the user's understanding and trust in the recommendation logic, reduce doubts or resistance caused by information asymmetry, and improve the user experience.

[0061] In a preferred embodiment, such as Figure 7 As shown, after resolving recommendation conflicts, based on the comprehensive score... The numerical values ​​are used to execute a three-level output grading algorithm to generate the final recommendation result label. The three-level grading includes three levels: "Recommended to eat", "Try in small amounts", and "Avoid eating", with each level corresponding to a specific comprehensive score threshold range.

[0062] When the overall score A score of ≥1.0 indicates that the food is classified as "recommended for consumption." This threshold setting signifies that the system will only provide a positive recommendation if the overall evaluation after conflict arbitration maintains a relatively clear positive bias. The recommendation reason generation module for this level will output natural language text such as, "After comprehensively considering your physical condition and disease management needs, this food is beneficial to your health and is recommended for moderate consumption as part of your daily diet."

[0063] When the overall score When the value is between -1.0 and 1.0, the food is classified as belonging to the "small amount trial" level. This intermediate level is a key innovation of this invention; it applies to foods that exhibit a moderate conflict between two health dimensions, and whose evaluation tendency is not clearly defined after arbitration. This level reflects a principle of prudence, neither completely prohibiting users from consuming the food and thus excessively restricting their dietary choices, nor overly restricting them through the explicit qualifier "small amount," but rather prompting users to control their intake. The recommendation reason generation module will output specific guidance text such as, "This food is beneficial to you in some health dimensions, but requires attention in other dimensions. It is recommended that you try it in small amounts, pay attention to your body's response, and consume no more than 50 grams per serving."

[0064] When the overall score When the value is less than -1.0, the system classifies the food as belonging to the "avoid eating" level. This threshold setting indicates that when the overall evaluation after arbitration still shows a clear negative tendency, the system will provide restrictive advice to ensure user safety. The recommendation reason generation module will output warning text such as "Considering your health condition, this food may be detrimental to your disease control or physical conditioning. It is recommended to avoid eating it or to consume it cautiously only under the guidance of a professional."

[0065] Furthermore, more detailed explanations of the recommendation reasons can be generated. This explanation uses natural language generation technology to transform key decision-making criteria from the conflict arbitration process into easily understandable textual descriptions. Specifically, the process of generating explanatory text based on recommendation result tags includes the following steps: First, the information extraction stage, where the system extracts key decision-making elements from the conflict arbitration process, including the recommendation tag category, the dominant health goal type, the conflict dimension identification results, the score values ​​for each dimension, and the final arbitration strategy type; Second, the template matching stage, where the corresponding text generation template is selected based on the recommendation tag category, maintaining different expression frameworks for the three tags: recommended consumption, consumption with caution, and not recommended consumption. Each framework includes three components: an introductory statement, a rationale, and a recommendation explanation; Third, the content filling stage, where the extracted decision-making elements are processed according to predefined semantic rules. The system fills in the template; for example, when the recommendation tag is "consume with caution" and the conflict type is "conflict between constitution and disease," the system will generate a statement like, "This food meets your nutritional needs, but considering the conflict between your constitution and disease management goals, it is recommended to consume it in moderation." The fourth step is the personalization adjustment stage, which adjusts the expression based on the user's historical interaction preferences. For users who prefer detailed explanations, specific numerical data and explanations from Traditional Chinese Medicine are provided; for users who prefer concise expressions, a simplified version is offered. The fifth step is the language optimization stage, which uses a pre-trained language model to optimize the fluency of the generated text, ensuring natural expression, logical coherence, and adherence to Chinese expression habits. Through this series of processes, the complex, multi-dimensional decision-making process can be transformed into clear, accurate, and easy-to-understand recommendations, helping users understand the health logic behind the recommendations. The rationale typically includes three elements: an evaluation and reason from a Traditional Chinese Medicine perspective (e.g., "This food is warming in nature and helps improve your Yang deficiency constitution"), an evaluation and reason from a nutritional perspective (e.g., "However, this food is high in fat, which is not conducive to your diabetes and blood lipid control"), and an explanation of the system's arbitration logic (e.g., "Considering your urgent need for disease control, the system recommends prioritizing disease management, therefore a small amount is suggested"). This transparent rationale helps improve the user's understanding and trust in the decision-making process of this embodiment.

[0066] In one possible implementation, the step of weighted summing of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generating a second dietary recommendation result for the target food based on the weighted summation result, includes: Based on the disease severity index and the constitution deviation index, the TCM constitution suitability score and the disease nutritional suitability score are weighted and summed to obtain a comprehensive score. An initial second dietary recommendation for the target food is generated based on the numerical range to which the comprehensive score belongs; If the disease status field in the confirmed disease list and the user's user information meet the preset conditions, then the preset priority rule or correction coefficient is invoked to correct the initial second dietary recommendation result, and the second dietary recommendation result is generated; otherwise, the initial second dietary recommendation result is used as the second dietary recommendation result.

[0067] like Figure 8 As shown, this embodiment further considers special scenario processing and result output. Before outputting the final recommendation result, special scenarios need to be identified and processed to ensure the safety and rationality of the recommendation result under various complex situations. Special scenario processing includes three sub-modules: multi-disease priority adjustment, differentiated treatment of acute and chronic diseases, and correction for special populations.

[0068] For users with multiple illnesses, although a comprehensive disease-based nutritional suitability score has been calculated using a weighted average in step S3, further adjustments to the decision-making logic are necessary in certain situations. The system maintains a disease priority rule base, defining the priority relationships between different disease categories. For example, when a user has both acute and chronic illnesses, the management needs for the acute illness have higher priority. The system checks if any disease marked "acute phase" exists in the user's disease list. If so, it applies a 1.5-fold weighting enhancement coefficient to the nutritional suitability score related to the acute illness in the comprehensive score calculation, ensuring that recommendations prioritize the treatment needs of the acute illness.

[0069] For specific population groups, the system applies corresponding correction strategies based on users' demographic characteristics (including age, gender, and physiological status). The system first extracts the age field from user profile data. If the user is under 18 years old, they are labeled as a child or adolescent. If the user is female and the demographic data includes the tags "pregnant" or "lactating," they are labeled as a pregnant or postpartum woman. For children and adolescents, the system reduces the restriction intensity on certain warming and drying foods when calculating the TCM constitution suitability score, because children have stronger Yang energy and relatively better tolerance to warming foods; the correction coefficient is set to 0.8. For pregnant or postpartum women, the system applies additional negative score correction to foods labeled as having "blood-activating and stasis-removing" or "cold" effects in TCM theory, with a correction coefficient of -1.5, to ensure the safety of pregnant women and fetuses. For the elderly population over 65 years old, the system applies a 1.2-fold enhancement coefficient when quantifying disease severity, because the elderly population has a relatively lower tolerance for health risks, and the safety requirements for disease management are more prominent.

[0070] After completing all special scenario processing, the system encapsulates the final recommendation results in a structured format. The output data structure includes recommendation result tags (suggested consumption / try in small amounts / avoid consumption), a comprehensive score, original scores for Traditional Chinese Medicine and disease dimensions, dynamic weighting coefficients, recommendation reason text, and processing timestamps. The system passes this structured output to the recommendation generation agent through a standardized data interface, which is responsible for integrating the recommendation results into the complete user interface. The entire conflict resolution process is now complete, and the system returns to a waiting state, ready to process the next food recommendation request.

[0071] Example 2: Example 2 illustrates the execution process and decision-making logic of the technical solution provided in this application in a practical application through a specific clinical scenario.

[0072] Case Scenario: User Zhang, male, 45 years old, has been diagnosed with type 2 diabetes for 3 years and hyperlipidemia for 2 years, both currently under medication control. His most recent physical examination showed: fasting blood glucose 8.2 mmol / L, HbA1c 7.8%, blood pressure 125 / 82 mmHg, total cholesterol 6.5 mmol / L (normal upper limit 5.2 mmol / L), LDL cholesterol 4.0 mmol / L (normal upper limit 3.4 mmol / L), and triglycerides 2.8 mmol / L (normal upper limit 1.7 mmol / L). After a standardized Traditional Chinese Medicine (TCM) constitution assessment, Zhang was diagnosed with Yang deficiency, scoring 78 points, indicating a relatively pronounced but not extreme Yang deficiency. Zhang took a photo of a dish using a mobile application. The system, through image recognition and food analysis, identified lamb as the main ingredient. The system now needs to perform a recommendation conflict resolution analysis for lamb.

[0073] S101: User Health Data Acquisition and Structured Processing The system extracted Zhang's health record information from the user profile database. The TCM constitution type field was read as "YX" (Yang deficiency constitution code), with a constitution assessment score of 78. The disease list field contained two records: the first record had an ICD-10 code of E11 (type 2 diabetes), diagnosed in May 2022, currently under "medication control," and the severity level field was temporarily empty and needed to be calculated later. The second record had an ICD-10 code of E78.5 (hyperlipidemia), diagnosed in March 2023, currently under "medication control," and the severity level field also needed to be calculated later. Physiological indicator data was extracted from the most recent physical examination record (timestamp 2025-10-01): FPG=8.2, HbA1c=7.8, blood pressure=125 / 82 mmHg, total cholesterol=6.5 mmol / L, LDL cholesterol=4.0 mmol / L, triglycerides=2.8 mmol / L.

[0074] The system receives the food identifier "FOOD_20231215_Lamb" and retrieves its complete attribute information from the knowledge base. Nutritional attributes show that per 100 grams of lamb, it contains 203 kcal of energy, 19.0g of protein, 14.1g of fat (including 5.2g of saturated fat), 0g of carbohydrates, 0g of dietary fiber, 92mg of sodium, and a glycemic index (GI) of 0 (pure protein food). Traditional Chinese medicine attributes show that it is warm in nature, sweet in taste, and enters the spleen and kidney meridians, with functions of warming the middle jiao and tonifying deficiency, tonifying the kidneys and assisting yang, and nourishing qi and blood. The system loads all data into its working memory and creates a data processing workspace, ready for scoring calculation. The system also records the processing start timestamp 2025-10-10T14:32:00Z for subsequent result output.

[0075] S201: Calculation of Traditional Chinese Medicine Constitution Compatibility Score The system retrieved dietary rules related to Yang deficiency constitution from the Traditional Chinese Medicine (TCM) dietary therapy rule base. Three matching rules relevant to the current scenario were identified. The first rule's antecedent was "User constitution = Yang deficiency AND food nature = warm," and its consequent was "suitability = high CONFIDENCE = 0.95." This rule was successfully matched because mutton is warm in nature. The system assigned a base score of 2.0 based on the "high" suitability level, multiplied by a confidence coefficient of 0.95, resulting in a weighted contribution value of 1.90. The second rule's antecedent was "User constitution = Yang deficiency AND food meridian tropism = Kidney meridian," and its consequent was "suitability = high CONFIDENCE = 0.90." This rule was also successfully matched because mutton tropisms the Kidney meridian. The weighted contribution value was calculated as 2.0 multiplied by 0.90, resulting in a value of 1.80. The antecedent of rule three is "User constitution = Yang deficiency AND food taste = sweet", and the consequent is "fitness = medium CONFIDENCE = 0.85". This rule matches because mutton tastes sweet, with a base score of 1.0 and a weighted contribution of 1.0 × 0.85 = 0.85.

[0076] The system sums the weighted contribution values ​​of the three matching rules, resulting in a total of 1.90 + 1.80 + 0.85 = 4.55. With a total of 3 matching rules, the system calculates a preliminary TCM constitution suitability score, which is 4.55 / 3 = 1.517. The system then executes a meridian and susceptible organ matching correction procedure. It queries the associated organ data for Yang deficiency constitution from the constitution-organ correlation matrix. The search results show that the correlation strength coefficient between Yang deficiency constitution and the spleen is 0.80, and the correlation strength coefficient with the kidney is 0.90. The meridian attributes of mutton include the spleen and kidney meridians, which highly match the two main susceptible organs of Yang deficiency constitution users. Based on the matching results, the system calculates a correction factor, taking the average of the two correlation strength coefficients (0.85) and multiplying it by a preset correction factor of 0.5, resulting in a positive correction factor of 0.425. The system adds the correction factor to the initial score, and calculates the corrected TCM constitution suitability score as 1.517 + 0.425 = 1.942.

[0077] The system performed a normalization check, confirming that the score of 1.942 was within the specified range of [-5, +5] and therefore no truncation was required. The final TCM constitution suitability score was then calculated. The score was determined to be 1.94 (rounded to two decimal places). This positive score indicates that mutton, from the perspective of traditional Chinese medicine theory, has a good conditioning effect on Mr. Zhang's Yang deficiency constitution, which is in line with the basic principle of "Yang deficiency should be treated with warming and tonifying" in traditional Chinese medicine. The system writes the score result to the temporary data storage area and records the timestamp of the score calculation completion in the processing log.

[0078] S301: Calculation of Nutritional Adequacy Score for Disease The system calculates the nutritional suitability scores for mutton based on the two diseases suffered by Mr. Zhang, and then performs weighted aggregation. The system first processes the nutritional suitability assessment for type 2 diabetes. It retrieves the nutritional intervention rule set corresponding to ICD-10 code E11 from the disease nutrition management knowledge base, extracting core rules including controlling carbohydrate intake, limiting saturated fat intake, prioritizing low glycemic index foods, and consuming adequate amounts of high-quality protein.

[0079] The system evaluated each nutritional component of lamb item by item. In terms of carbohydrates, lamb contains 0 grams of carbohydrates, far below the recommended daily intake for diabetic patients (typically 45-60 grams). This dimension was rated highly favorable, and the system assigned a positive score of 1.0. In terms of glycemic index, lamb, as a pure protein food, has a GI value of 0, having minimal impact on postprandial blood glucose. This dimension was also rated highly favorable, and the system assigned a positive score of 1.5. In terms of protein, lamb contains 19.0 grams of high-quality protein per 100 grams, which is beneficial for diabetic patients in maintaining muscle mass and nutritional status. This dimension was rated favorable, and the system assigned a positive score of 1.0. In the saturated fat dimension, lamb contains 5.2 grams of saturated fat per 100 grams. The system retrieved the recommended limit for saturated fat intake per serving for diabetic patients from the rule base, which is 3.0 grams per 100 grams. The actual content is 5.2 / 3.0 = 1.73, exceeding the limit. Therefore, this dimension is assessed as unfavorable, and the system assigns a negative score of -1.5. Based on the excess ratio of 1.73, an excess penalty coefficient is applied, resulting in a final score of -1.5 × 1.73 = -2.60 for this dimension.

[0080] The system assigns weights to each dimension based on their clinical importance in diabetes nutrition management. The glycemic index (GI), a core indicator of diabetes dietary management, has a weight of 0.35. Carbohydrate control is equally important, with a weight of 0.25. Saturated fat restriction is crucial for cardiovascular risk management in diabetic patients, with a weight of 0.25. Protein intake is relatively less important, with a weight of 0.15. The system performs a weighted summation calculation, and the single-disease suitability score for lamb in type 2 diabetes is: 1.5 × 0.35 + 1.0 × 0.25 + (-2.60 × 0.25) + 1.0 × 0.15 = 0.525 + 0.25 -0.65 + 0.15 = 0.275.

[0081] The system then processes the nutritional suitability assessment for hyperlipidemia. It retrieves the nutritional intervention rule set corresponding to ICD-10 code E78.5 from the disease nutrition management knowledge base, extracting core rules including strictly limiting saturated fat intake, controlling total fat intake, reducing cholesterol intake, and increasing dietary fiber intake. Nutritional management for hyperlipidemia requires stricter restrictions on fat components than for diabetes.

[0082] The system performed a nutritional assessment of mutton from the perspective of hyperlipidemia. In the saturated fat dimension, the recommended limit for hyperlipidemia patients is lower, at 2.0 g / 100g. The actual fat content of mutton (5.2g) exceeds this limit by 2.6, resulting in a significantly higher limit. This dimension is assigned a negative score of -2.0. After applying an excess penalty coefficient of 2.6, the final score is -2.0 × 2.6 = -5.20. In the total fat dimension, mutton contains 14.1g / 100g, exceeding the recommended single-serving total fat limit of 10.0g / 100g for hyperlipidemia patients, with an excess rate of 1.41. This dimension is assigned a negative score of -1.5 × 1.41 = -2.12. In the dietary fiber dimension, mutton, as an animal-based food, contains no dietary fiber, while dietary fiber is beneficial for lowering blood lipids. This dimension is assessed as lacking a beneficial factor, assigning a slightly negative score of -0.5. In terms of protein, although mutton is rich in high-quality protein, in the nutritional management priorities for hyperlipidemia, the positive effects of protein are overshadowed by the negative effects of fat content, and this dimension is assigned only a neutral score of 0.

[0083] The system assigns weight coefficients to each dimension of nutritional management for hyperlipidemia. Saturated fat, as the most critical management target, has a weight of 0.45. Total fat control is the next most important, with a weight of 0.30. The promoting effect of dietary fiber intake has a weight of 0.15. Protein has a weight of 0.10. The system performs a weighted summation calculation, and the single-disease suitability score of mutton for hyperlipidemia is (-5.20 × 0.45) + (-2.12 × 0.30) + (-0.5 × 0.15) + 0 × 0.10 = -2.34 - 0.636 - 0.075 + 0 = -3.05.

[0084] The system now needs to aggregate two individual disease scores into a comprehensive disease-nutrition suitability score. The system invokes the disease severity assessment module to calculate a severity score for each disease based on Zhang's disease type and physiological indicators. For type 2 diabetes, the system retrieves a baseline risk score of 3 from the disease risk grading table. The system further assesses the disease control status. Zhang's fasting blood glucose of 8.2 mmol / L and glycated hemoglobin of 7.8% are both on the edge of the target control range, and the control status is judged as "moderate," with a corresponding control status correction coefficient of 1.0. The severity score for diabetes is calculated as 3 × 1.0 = 3.0. For hyperlipidemia, the system retrieves a baseline risk score of 2. Zhang's blood lipid indicators show total cholesterol of 6.5 mmol / L (exceeding the standard by 25%), low-density lipoprotein of 4.0 mmol / L (exceeding the standard by 18%), and triglycerides of 2.8 mmol / L (exceeding the standard by 65%). Multiple indicators are significantly above the standard, and the control status is judged as "poor," with a corresponding control status correction coefficient of 1.3. The severity score for hyperlipidemia is calculated as 2 × 1.3 = 2.6.

[0085] The system normalizes the severity scores of the two diseases to calculate the aggregate weight. The severity score of diabetes (3.0) and the severity score of hyperlipidemia (2.6) are added together to form a total of 5.6. The normalized weight for diabetes is 3.0 / 5.6 = 0.536. The normalized weight for hyperlipidemia is 2.6 / 5.6 = 0.464. The system then performs a weighted average calculation to integrate the nutritional suitability scores of the two diseases. The result is 0.275 × 0.536 + (-3.05 × 0.464) = 0.147 - 1.415 = -1.27. This negative score indicates that mutton is generally detrimental to Zhang's current health from a disease nutrition management perspective, mainly because its high fat content conflicts with the nutritional management requirements for hyperlipidemia.

[0086] S401: Recommendation Conflict Identification and Judgment The system executes a conflict identification algorithm, using the two pre-calculated suitability scores for judgment. Current TCM constitution suitability score. The value was 1.94, indicating a disease-related nutritional suitability score. The value is -1.27. The system first checks the sign opposite condition, calculating the product of the two scores as 1.94 × 1.27 = -2.46, which is less than 0, satisfying the sign opposite condition. This indicates that mutton received a positive evaluation in the TCM constitution conditioning dimension, but a negative evaluation in the disease nutrition management dimension. The recommendations for the two health dimensions are opposite, indicating a potential conflict.

[0087] The system then checks the significance threshold. The absolute value of the TCM constitution suitability score is 1.94, which is greater than the preset significance threshold of 1.5, thus meeting the significance requirement. The absolute value of the disease-nutritional suitability score is 1.27, which is less than the significance threshold of 1.5, thus not meeting the significance requirement. According to strict conflict resolution logic, only when the absolute values ​​of both scores reach the significance threshold is a substantial conflict deemed to require arbitration. However, considering... The absolute value of 1.27 is close to the threshold of 1.5, and Zhang suffers from two chronic diseases closely related to diet, so the system's security protection mechanism is triggered.

[0088] The system incorporates a safety boundary determination logic within its conflict identification module. When the disease-nutrition suitability score is negative and its absolute value is greater than 1.0, even if the significance threshold of 1.5 is not fully met, the conflict arbitration process will still be initiated if one of the following conditions is simultaneously met: Condition 1: The user suffers from two or more diseases closely related to nutrition management. Condition 2: The user's key physiological indicators show significant deviations. Zhang's case meets both conditions; therefore, the system determines that the scenario carries sufficient clinical risk and decides to initiate the conflict arbitration process to ensure the safety and rationality of the recommendation results. The system records the conflict determination result in the processing log as "Recommendation conflict exists, arbitration process triggered, determination timestamp 2025-10-10T14:32:08Z".

[0089] S501: Quantification of Disease Severity and Calculation of Physical Deviation The system enters the conflict arbitration process, first calculating two key parameters used for dynamic weight adjustment. The first parameter is a comprehensive disease severity quantification index. In the previous step S3, the system had already calculated the severity scores for Zhang's two diseases: a severity score of 3.0 for type 2 diabetes and a severity score of 2.6 for hyperlipidemia. Since dynamic weight calculation requires a single value that can represent the user's overall disease burden, the system uses a weighted average method to calculate the comprehensive severity index.

[0090] The system uses the same weighting coefficients as the disease score aggregation, i.e., diabetes weighted at 0.536 and hyperlipidemia weighted at 0.464, thus maintaining consistency in the assessment. This is a comprehensive quantitative indicator of disease severity. The calculation is 3.0 × 0.536 + 2.6 × 0.464 = 1.608 + 1.206 = 2.81. This value indicates that Mr. Zhang's current disease burden is at a moderately high level, and the need for safety in disease management is relatively prominent. The system stores this value in the arbitration calculation workspace.

[0091] The system then calculates the second parameter, namely the physical fitness deviation index. This indicator is used to quantify the degree of deviation between a user's current physical condition and their ideal health state. The system extracts Zhang's TCM constitution type, "Yang Deficiency," and constitution assessment score of 78 from the user profile data. The system queries the constitution deviation mapping table; Yang Deficiency, as a common imbalanced constitution, has a base deviation value set at 2, belonging to the mild imbalanced constitution category.

[0092] The system further adjusts the baseline deviation based on the physical fitness assessment score. A physical fitness assessment score of 78 points falls within the 60-80 point range, indicating that Zhang's Yang deficiency constitution is relatively pronounced but not extremely typical. According to the adjustment rules, the adjustment strategy for this score range is to maintain the baseline deviation unchanged, without applying any enhancement or weakening coefficients. Therefore, the final physical fitness deviation index... The value is set at 2.0. This indicates that Zhang's physical imbalance is mild, and while dietary adjustments to improve his constitution are still necessary, they are not as urgent as disease management. The system stores this value in the arbitration calculation workspace, preparing for the next step of dynamic weight calculation.

[0093] S601: Dynamic Weight Calculation and Comprehensive Score Generation The system is based on quantitative indicators of disease severity. Equal to 2.81 and the physical deviation index The value is set to 2.0, triggering the dynamic weighting calculation algorithm. The system first calculates the importance factors for the two health dimensions. The importance factor for the nutrition dimension is also calculated. The calculation follows the formula: the baseline coefficient α multiplied by the disease severity index, where the current value of α is 1.5. The result is 1.5 × 2.81 = 4.215. This value reflects the priority and intensity of nutritional management required for Zhang's current disease condition.

[0094] The system then calculates the importance factors for the Traditional Chinese Medicine dimension. The calculation, based on the formula β multiplied by the constitution deviation index, with β currently set at 1.0, yields a result of 1.0 × 2.0 = 2.0. This value reflects the priority of Mr. Zhang's need for constitution conditioning. A comparison of the two importance factors shows that the nutritional dimension is significantly more important than the traditional Chinese medicine dimension, implying that disease management should receive a higher weighting under the current circumstances.

[0095] The system performs normalization, converting importance factors into weight coefficients. The sum of the two importance factors is 4.215 + 2.0 = 6.215. (Weight coefficients for the nutritional dimension) The calculation is 4.215 / 6.215 = 0.678. This is the weighting coefficient for the Traditional Chinese Medicine dimension. The calculation is 2.0 / 6.215 = 0.322. The system verifies that the sum of the two weight coefficients is 0.678 + 0.322 = 1.0, which meets the normalization requirement. This weight allocation result indicates that in conflict arbitration, considerations of disease nutrition management will account for approximately 68% of the weight, while considerations of traditional Chinese medicine constitution conditioning will account for approximately 32% of the weight.

[0096] The system uses the calculated dynamic weighting coefficients to perform a weighted average of the two original fit scores to generate the final comprehensive score. The calculation formula is: Substituting the values, we get 0.322 × 1.94 + 0.678 × (-1.27) = 0.625 - 0.861 = -0.236. The negative overall score indicates that after intelligent arbitration and weight adjustment, the negative evaluation of disease nutrition management dominated the overall recommendation tendency, although the positive evaluation of the Traditional Chinese Medicine constitution conditioning dimension provided some offsetting effect. The system will use the overall score... Write -0.24 (rounded to two decimal places) into the arbitration result data structure and record the timestamp of weight calculation completion in the processing log.

[0097] S701: Three-level output classification determination The system is based on a comprehensive score. For a value equal to -0.24, a three-level output grading algorithm is executed. The system first checks if the score is greater than or equal to 1.0; if not, the "recommended to eat" level is excluded. The system then checks if the score is less than -1.0; if not, the "avoid eating" level is excluded. If the system confirms that the overall score is within the range of -1.0 to 1.0, according to the grading rules, the food should be classified as "try in small amounts" level.

[0098] The system invokes the recommendation reason generation module, which uses natural language generation technology to generate user-friendly explanatory text based on key data from the arbitration process. The reason generation module first extracts evaluation information from the Traditional Chinese Medicine dimension, including... The value is 1.94, which relates to the traditional Chinese medicine properties of mutton (warm in nature, sweet in taste, and associated with the spleen and kidney meridians), and the conditioning needs of those with a Yang deficiency constitution. The module generates the first part of the text: "Mutton is warm in nature and sweet in taste, and associated with the spleen and kidney meridians. From a traditional Chinese medicine perspective, it helps improve your Yang deficiency constitution, and has the effects of warming the middle and tonifying deficiency, and tonifying the kidneys and assisting Yang, which is beneficial to your constitution conditioning." The reason generation module then extracts evaluation information from the disease's nutritional dimensions, including... The value is -1.27, the nutritional components of mutton (especially saturated fat 5.2g / 100g), and Mr. Zhang's two diseases and high blood lipid levels. The module generates the second part of the text: "However, mutton has a high saturated fat content, containing 5.2g of saturated fat per 100g. Considering that you currently have both type 2 diabetes and hyperlipidemia, and your blood lipid levels are poorly controlled, with total cholesterol and LDL cholesterol significantly exceeding the standard, excessive intake of mutton may aggravate the blood lipid burden and is detrimental to disease management and cardiovascular health." The reason generation module finally extracts the arbitration decision information, including the dynamic weight allocation result (disease management weight 68%, physical conditioning weight 32%) and the final comprehensive score of -0.24, generating the third part of the text: "Considering your health condition and the relatively more urgent need for disease control safety, the system suggests that you can try a small amount of mutton to achieve some physical conditioning effects, but you must strictly control the amount and frequency of consumption. It is recommended to eat it no more than once a week, with each serving limited to less than 50 grams (about one tael), and to prioritize lean meat, removing visible fat. It is recommended to eat it with vegetables rich in dietary fiber, and to monitor postprandial blood sugar after consumption and regularly check blood lipid levels. If blood lipid levels remain below the target, it is recommended to stop eating it and consult your attending physician or nutritionist." The system integrates the three parts of text to form a complete explanation of the recommendation, totaling approximately 260 words. This provides both a transparent explanation of the decision-making basis and specific, actionable suggestions. This detailed explanation helps improve users' understanding and trust in the system's recommendations, while also providing ample information support for their dietary decisions.

[0099] S801: Special Scenario Handling and Result Output Before outputting the final recommendation results, the system executes a special scenario processing procedure to ensure the safety and rationality of the recommendations under various complex situations. The system first initiates a multi-disease priority check submodule. This submodule extracts two diseases for Mr. Zhang from the user's disease list: type 2 diabetes (ICD-10 code E11) and hyperlipidemia (ICD-10 code E78.5). The system queries the disease priority rule base to check if there is a special priority relationship between the two diseases. The results returned by the rule base show that both diseases are chronic metabolic diseases, and there is no situation where one disease is significantly prioritized over the other. The system further checks the disease status field, confirming that neither disease is in an acute phase or acute exacerbation state; therefore, there is no need to apply an acute disease priority adjustment strategy. The conclusion of this submodule is that no priority correction is needed for the current recommendation results.

[0100] The system then activated the special population identification and correction submodule. This submodule extracted demographic information from the user profile data. The age field showed that Zhang was 45 years old, neither belonging to the child / adolescent group (defined as under 18 years old) nor the elderly group (defined as over 65 years old). The gender field showed male, so there was no need to check for pregnancy or breastfeeding status. The system further checked the user profile for special health status labels, such as immunosuppression, renal insufficiency, liver insufficiency, etc., which require special attention. The search result was no special labels. The submodule concluded that the user did not belong to any special population requiring correction strategies, and the recommendation result remained unchanged.

[0101] After completing all special scenario processing procedures, the system confirms that the final recommendation result requires no adjustment. The system then encapsulates the recommendation result in a structured manner, creating an output data object. This object contains the following fields: the recommendation level field is assigned the value "Minor Trial", the comprehensive score field is assigned the value -0.24, the TCM constitution suitability score field is assigned the value 1.94, the disease nutrition suitability score field is assigned the value -1.27, the TCM dimension weight field is assigned the value 0.322, the disease dimension weight field is assigned the value 0.678, the recommendation reason field is assigned the previously generated complete explanatory text, the processing timestamp field is assigned the value 2025-10-10T14:32:15Z, the user identifier field is assigned the user ID of Zhang, and the food identifier field is assigned the value "FOOD_20231215_Lamb".

[0102] The system transmits encapsulated output data objects to the recommendation-generating agent via a standardized Application Programming Interface (API). This interface uses JSON (JavaScript Object Notation) data format for exchange, ensuring structured and reliable data transmission. Upon receiving the data, the recommendation-generating agent integrates the recommendations into a complete user interface, including generating visual rating comparison charts, highlighting key information, and providing further dietary alternative suggestions. After completing the data transmission, the system cleans up temporary computational data in its working memory, archives the processing results to a historical database for future user behavior analysis and system optimization, and then returns to standby mode to prepare for receiving and processing the next food recommendation request.

[0103] Compared with existing technologies, the embodiments of this application have the following significant advantages. First, the embodiments of this application fill the technical gap in resolving conflicts across theoretical systems in the field of integrative medicine, and for the first time realize the quantitative and automated arbitration of the contradictions between traditional Chinese medicine constitution theory and modern disease nutrition, providing users with unified and clear dietary guidance and avoiding confusion and decision-making burden when faced with conflicting suggestions. Second, the embodiments of this application achieve truly personalized recommendations through dynamic weight adjustment and multi-parameter comprehensive decision-making, enabling the system to flexibly adjust the decision logic according to each user's specific health condition, significantly improving its adaptability and accuracy compared to recommendation systems with fixed rules. Third, the three-level output hierarchy system of the embodiments of this application, while ensuring security, retains more dietary choices for users, avoiding the excessive restrictions that may be caused by traditional binary judgments, and improving the practicality of the recommendation results and user acceptance. Finally, the processing strategies of the embodiments of this application for complex scenarios such as multiple coexisting diseases and special populations enhance the robustness and reliability of the system in practical applications, enabling it to meet the diverse needs of clinical practice and possessing broad application prospects and commercial value.

[0104] Example 3: like Figure 9 As shown, Embodiment 3 provides a diet recommendation system based on multi-dimensional data analysis, including an acquisition module 10, a first matching module 20, a second matching module 30, a conflict judgment module 40, an evaluation module 50, and a conflict resolution module 60. The acquisition module 10 is used to acquire and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base, based on the target food. The first matching module 20 is used to perform a first rule matching in a preset TCM dietary therapy rule library based on the TCM theoretical attribute set and the user's TCM constitution type, and calculate the compatibility score between the target food and the user's TCM constitution based on the first rule matching result; The second matching module 30 is used to perform second rule matching in a preset disease nutrition rule base according to the set of modern nutritional attributes and the user's list of diagnosed diseases, and calculate the disease nutrition suitability score between the target food and the user based on the second rule matching result; The conflict judgment module 40 is used to determine whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutrition suitability score. If there is no dietary recommendation conflict, a first dietary recommendation result for the target food is generated based on the TCM constitution suitability score and the disease nutrition suitability score. The assessment module 50 is used to assess the severity of the disease and the deviation of the constitution of the user based on the user's list of diagnosed diseases and TCM constitution type if there is a conflict in dietary recommendations, and to obtain the corresponding disease severity index and constitution deviation index. The conflict resolution module 60 is used to perform a weighted summation of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generate a second dietary recommendation result for the target food based on the weighted summation result.

[0105] Furthermore, the diet recommendation system also includes a recommendation text generation module. This module is used to generate corresponding recommendation text based on the second diet recommendation result when a second diet recommendation result for the target food is generated based on the weighted summation result, and then sends the second diet recommendation result and the recommendation text to a designated device. Specifically: Based on the TCM constitution suitability score and the user's TCM constitution type, generate the first evaluation text of the target food in the TCM dimension; Based on the disease nutritional suitability score and the user's list of diagnosed diseases, a second evaluation text of the target food in terms of nutrition is generated. Based on the disease severity index, the body constitution deviation index, and the second dietary recommendation result, an explanatory text for the dietary recommendation conflict is generated; The recommended text is generated by combining the first evaluation text, the second evaluation text, and the explanation text, and the second dietary recommendation result and the recommended text are sent to the designated device.

[0106] In one possible implementation, the step of weighted summing of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generating a second dietary recommendation result for the target food based on the weighted summation result, includes: Based on the disease severity index and the constitution deviation index, the TCM constitution suitability score and the disease nutritional suitability score are weighted and summed to obtain a comprehensive score. An initial second dietary recommendation for the target food is generated based on the numerical range to which the comprehensive score belongs; If the disease status field in the confirmed disease list and the user's user information meet the preset conditions, then the preset priority rule or correction coefficient is invoked to correct the initial second dietary recommendation result, and the second dietary recommendation result is generated; otherwise, the initial second dietary recommendation result is used as the second dietary recommendation result.

[0107] This application provides a dietary recommendation system based on multi-dimensional data analysis. By comprehensively considering both traditional Chinese medicine (TCM) theory and modern nutrition standards, and combining multi-parameter comprehensive decision-making techniques, it achieves truly personalized recommendations, thereby effectively improving the accuracy and practicality of dietary recommendations. In modern society, people's health conditions are increasingly complex, often suffering from certain specific diseases while possessing a particular TCM constitution. In such cases, relying solely on a single theoretical system may not fully meet the needs, and may even produce contradictory recommendation results. This embodiment introduces a TCM constitution suitability score and a disease-nutritional suitability score, which not only objectively reflects the suitability of food in different dimensions but also dynamically detects potential recommendation conflicts, ensuring that the final dietary recommendations are more scientific and reasonable. Furthermore, this embodiment also introduces disease severity assessment and constitution deviation assessment mechanisms, which can flexibly adjust weight allocation when dietary recommendation conflicts occur, determine the focus of dietary recommendations based on the user's own situation, and resolve conflicts through weighted summation to obtain the final recommendation result. Compared to traditional single-theory recommendation systems, this embodiment significantly enhances the personalization and comprehensiveness of health management, helping users make optimal choices in complex health states, and also provides an important reference for future smart healthcare and big data-driven health interventions.

[0108] For a more detailed explanation of the working principle and procedures of this embodiment, please refer to the relevant description in Embodiment 1.

[0109] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of this application. It should be understood that the above descriptions are merely specific embodiments of this application and are not intended to limit the scope of protection of this application. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application for those skilled in the art.

Claims

1. A dietary recommendation method based on multi-dimensional data analysis, characterized in that, include: Acquire and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base; Based on the set of TCM theoretical attributes and the user's TCM constitution type, a first rule matching is performed in the preset TCM dietary therapy rule library, and the compatibility score between the target food and the user's TCM constitution is calculated based on the first rule matching result. Based on the set of modern nutritional attributes and the user's list of diagnosed diseases, a second rule matching is performed in a preset disease nutrition rule base, and the disease nutrition suitability score between the target food and the user is calculated based on the second rule matching result. Based on the TCM constitution suitability score and the disease nutrition suitability score, it is determined whether there is a conflict in dietary recommendations. If there is no conflict in dietary recommendations, a first dietary recommendation result for the target food is generated based on the TCM constitution suitability score and the disease nutrition suitability score. If there is a conflict in dietary recommendations, the severity of the disease and the deviation of the constitution will be assessed based on the user's list of diagnosed diseases and TCM constitution type, respectively, to obtain the corresponding disease severity index and constitution deviation index. Based on the disease severity index and the body constitution deviation index, the TCM body constitution suitability score and the disease nutritional suitability score are weighted and summed, and a second dietary recommendation result for the target food is generated based on the weighted summation result.

2. The dietary recommendation method based on multi-dimensional data analysis as described in claim 1, characterized in that, The process of acquiring and extracting the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base includes: The system acquires text data or image data and inputs it into a preset food analysis agent, so that the food analysis agent can call the corresponding image recognition algorithm or text parsing algorithm to determine the corresponding target food based on the text data or image data. Based on the target food, extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base; The set of modern nutritional properties includes the energy value per unit mass of food, protein content, fat content, carbohydrate content, dietary fiber content, glycemic index, and the content of various vitamins; The set of TCM theoretical attributes includes the nature, flavor, meridian tropism, and efficacy description of the target food.

3. The dietary recommendation method based on multi-dimensional data analysis as described in claim 1, characterized in that, The process involves performing a first rule matching within a pre-defined TCM dietary therapy rule base based on the TCM theoretical attribute set and the user's TCM constitution type. The result of the first rule matching is used to calculate the compatibility score between the target food and the user's TCM constitution, including: Based on the set of TCM theoretical attributes and the user's TCM constitution type, a search is performed in the TCM dietary therapy rule base to obtain several TCM constitution suitability rules. Each TCM constitution suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any TCM theoretical attribute and any TCM constitution type, and the consequent conclusions are the suitability level and confidence level of the TCM theoretical attribute and TCM constitution type in the antecedent conditions. Based on the consequent conclusions in each of the aforementioned TCM constitution suitability rules, the TCM constitution suitability score between the target food and the user is calculated.

4. The dietary recommendation method based on multi-dimensional data analysis as described in claim 3, characterized in that, The step of calculating the TCM constitution compatibility score between the target food and the user based on the consequent conclusions in each of the TCM constitution compatibility rules includes: The fitness level and confidence level in the consequent conclusion of each of the TCM constitution suitability rules are multiplied to obtain the corresponding rule fitness score. The initial TCM constitution suitability score is obtained by summing the suitability scores of each rule. Based on the user's TCM constitution type and the preset constitution-organ correlation matrix, determine the correlation strength between the user's TCM constitution type and each organ. Based on the user's TCM constitution type and the correlation strength between each organ and the food meridian attribute in the TCM theory attribute set, the degree of matching between the target food and the user's organs is determined. The initial TCM constitution compatibility score is corrected based on the degree of matching between the target food and the user's internal organs, and the corrected score is normalized to a preset numerical range to obtain the TCM constitution compatibility score.

5. The dietary recommendation method based on multi-dimensional data analysis as described in claim 1, characterized in that, The step involves performing a second rule matching based on the set of modern nutritional attributes and the user's list of diagnosed diseases in a preset disease nutrition rule base, and calculating the disease nutrition suitability score between the target food and the user based on the second rule matching result, including: Iterate through each confirmed disease in the list of confirmed diseases. For any confirmed disease, search the disease nutrition rule base according to the confirmed disease and match the corresponding disease nutrition suitability rule. The disease nutrition suitability rule consists of antecedent conditions and consequent conclusions. The antecedent conditions include any disease name. The consequent conclusions include each recommended nutrient, the intake range of each recommended nutrient, each restricted nutrient, and the restricted intake range of each restricted nutrient. The set of modern nutritional attributes is compared with the nutritional suitability rules for each of the diseases, and the suitability score between the target food and each of the diagnosed diseases is determined based on the comparison results. The disease nutritional suitability score between the target food and the user is calculated based on each of the aforementioned suitability scores.

6. The dietary recommendation method based on multi-dimensional data analysis as described in claim 1, characterized in that, The step of determining whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutritional suitability score, and if no dietary recommendation conflict exists, then generating a first dietary recommendation result for the target food based on the TCM constitution suitability score and the disease nutritional suitability score, including: The dietary recommendation score is obtained by multiplying the TCM constitution suitability score and the disease nutrition suitability score. If the dietary recommendation score is less than 0 and the absolute values ​​of the TCM constitution suitability score and the disease nutrition suitability score are both greater than a preset threshold, then a dietary recommendation conflict is determined to exist; otherwise, no dietary recommendation conflict is determined to exist. When no dietary recommendation conflict is determined, if both the TCM constitution suitability score and the disease nutritional suitability score are positive, a positive dietary recommendation result for the target food is generated; if both the TCM constitution suitability score and the disease nutritional suitability score are negative, a negative dietary recommendation result for the target food is generated; if the signs of the TCM constitution suitability score and the disease nutritional suitability score are different, the generation of a positive or negative dietary recommendation result is determined based on the absolute values ​​of the TCM constitution suitability score and the disease nutritional suitability score.

7. The dietary recommendation method based on multi-dimensional data analysis as described in claim 1, characterized in that, If a conflict exists in dietary recommendations, the user's disease severity and constitution deviation will be assessed based on their confirmed disease list and TCM constitution type, respectively, to obtain corresponding disease severity and constitution deviation indicators, including: Based on the list of confirmed diseases and the preset disease risk classification table, determine the basic risk score for each confirmed disease in the list of confirmed diseases; Based on the user's physiological index data, the actual control status of each of the diagnosed diseases is evaluated, and the control status correction coefficient of each diagnosed disease is obtained. The severity index of the disease is calculated based on the basic risk score and control status correction coefficient of each diagnosed disease. The user's basic constitution deviation index is determined based on the TCM constitution type and the preset constitution deviation mapping table; Obtain the TCM constitution assessment form uploaded by the user, and determine the user's current constitution characteristic score based on the TCM constitution assessment form; The physical fitness deviation index is calculated based on the basic physical fitness deviation index and the current physical fitness characteristic score.

8. A dietary recommendation method based on multi-dimensional data analysis as described in any one of claims 1-7, characterized in that, The dietary recommendation method further includes, when generating a second dietary recommendation result for the target food based on the weighted summation result, generating corresponding recommendation text based on the second dietary recommendation result, and sending the second dietary recommendation result and the recommendation text to a designated device, specifically: Based on the TCM constitution suitability score and the user's TCM constitution type, generate the first evaluation text of the target food in the TCM dimension; Based on the disease nutritional suitability score and the user's list of diagnosed diseases, a second evaluation text of the target food in terms of nutrition is generated. Based on the disease severity index, the body constitution deviation index, and the second dietary recommendation result, an explanatory text for the dietary recommendation conflict is generated; The recommended text is generated by combining the first evaluation text, the second evaluation text, and the explanation text, and the second dietary recommendation result and the recommended text are sent to the designated device.

9. A dietary recommendation method based on multi-dimensional data analysis as described in any one of claims 1-7, characterized in that, The step involves weighting and summing the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generating a second dietary recommendation for the target food based on the weighted summation result, including: Based on the disease severity index and the constitution deviation index, the TCM constitution suitability score and the disease nutritional suitability score are weighted and summed to obtain a comprehensive score. An initial second dietary recommendation for the target food is generated based on the numerical range to which the comprehensive score belongs; If the disease status field in the confirmed disease list and the user's user information meet the preset conditions, then the preset priority rule or correction coefficient is invoked to correct the initial second dietary recommendation result, and the second dietary recommendation result is generated; otherwise, the initial second dietary recommendation result is used as the second dietary recommendation result.

10. A diet recommendation system based on multi-dimensional data analysis, characterized in that, It includes an acquisition module, a first matching module, a second matching module, a conflict determination module, an evaluation module, and a conflict resolution module; The acquisition module is used to acquire and extract the set of modern nutritional attributes and the set of traditional Chinese medicine theoretical attributes of the target food from a preset food knowledge base, based on the target food. The first matching module is used to perform a first rule matching in a preset TCM dietary therapy rule library based on the set of TCM theoretical attributes and the user's TCM constitution type, and calculate the compatibility score between the target food and the user's TCM constitution based on the first rule matching result; The second matching module is used to perform second rule matching in a preset disease nutrition rule base according to the set of modern nutritional attributes and the user's list of diagnosed diseases, and calculate the disease nutrition suitability score between the target food and the user based on the second rule matching result; The conflict judgment module is used to determine whether there is a dietary recommendation conflict based on the TCM constitution suitability score and the disease nutrition suitability score. If there is no dietary recommendation conflict, a first dietary recommendation result for the target food is generated based on the TCM constitution suitability score and the disease nutrition suitability score. The assessment module is used to assess the severity of the disease and the deviation of the constitution of the user based on the user's list of diagnosed diseases and TCM constitution type if there is a conflict in dietary recommendations, and to obtain the corresponding disease severity index and constitution deviation index. The conflict resolution module is used to perform a weighted summation of the TCM constitution suitability score and the disease nutritional suitability score based on the disease severity index and the constitution deviation index, and generate a second dietary recommendation result for the target food based on the weighted summation result.

Images