Coffee formula recommendation method and system based on user health data

Abstract

The invention provides a coffee formula recommendation method and system based on user health data, and relates to the technical field of data analysis, and the method comprises the steps: collecting the types of coffee blending elements, carrying out the interaction influence analysis of all types of coffee blending elements, and constructing an interaction influence map; reading user health data of the target user, wherein the user health data comprises physiological parameters, behavior records, disease history data and emotional state scores; performing hard component constraint and function preference feature extraction based on the user health data, and constructing a user health preference portrait; and performing formula adaptation analysis in combination with the interaction influence map and the user health preference portrait, and generating a recommended coffee formula for the target user. The technical problem that recommendation accuracy and user satisfaction are reduced due to the fact that the coffee formula recommendation method in the prior art depends on historical purchase records or taste preferences of the user to perform recommendation and cannot perform timely adjustment when the health condition or behavior habit of the user changes is solved.

Description

Technical Field

[0001] The present invention relates to the field of data analysis technology, and in particular to a method and system for recommending coffee recipes based on user health data. Background Art

[0002] With the improvement of health awareness, consumers' demand for coffee drinks has gradually shifted from a single taste and flavor to more functional and health-oriented products. This makes coffee recipe design more complex and personalized. However, most coffee recommendation systems on the market are still based on limited data such as user preferences and historical consumption records, lacking true health adaptation and functional optimization.

[0003] Specifically, most existing technologies use static health labels or questionnaires to analyze users' coffee needs. These data are generally obtained once and cannot track changes in users' health status in a timely manner. The recommendation system cannot make timely adjustments when users' health status or behavioral habits change. Especially with the continuous advancement of big data analysis technology, dynamic health data such as users' mood fluctuations and changes in exercise intensity can provide the system with more accurate user demand feedback. However, existing technologies have failed to make full use of these real-time updated health data for intelligent adjustments, resulting in the inability to update recipe recommendations in real time and achieve high personalization and dynamic adaptation. This not only reduces the accuracy of recommendations, but also affects user satisfaction and long-term stickiness, and cannot meet modern consumers' high expectations for healthy, personalized and intelligent services. Summary of the Invention

[0004] This application provides a coffee recipe recommendation method and system based on user health data, aiming to solve the technical problem that the coffee recipe recommendation method in the prior art relies on the user's historical purchase records or taste preferences for recommendation, and is unable to make timely adjustments when the user's health status or behavioral habits change, resulting in reduced recommendation accuracy and user satisfaction.

[0005] The first aspect disclosed in the present application provides a coffee recipe recommendation method based on user health data, the method comprising: collecting categories of coffee blending elements, performing interaction impact analysis on various types of coffee blending elements, and constructing an interaction impact map; reading user health data of a target user, including physiological parameters, behavior records, medical history data, and emotional state scores; performing hard component constraints and functional preference feature extraction based on the user health data to construct a user health preference portrait; performing recipe adaptation analysis in combination with the interaction impact map and the user health preference portrait to generate a recommended coffee recipe for the target user.

[0006] The second aspect disclosed in the present application provides a coffee recipe recommendation system based on user health data, which is used for the above-mentioned coffee recipe recommendation method based on user health data. The system includes: an interactive impact analysis module, which is used to collect coffee blending element categories, perform interactive impact analysis on various coffee blending elements, and construct an interactive impact map; a health data reading module, which is used to read the user health data of the target user, including physiological parameters, behavior records, medical history data and emotional state scores; a preference portrait construction module, which is used to perform hard component constraints and functional preference feature extraction based on the user health data to construct a user health preference portrait; a recipe adaptation analysis module, which is used to perform recipe adaptation analysis in combination with the interactive impact map and the user health preference portrait to generate a recommended coffee recipe for the target user.

[0007] One or more technical solutions provided in this application have at least the following beneficial effects:

[0008] By comprehensively collecting the categories of coffee blending elements and systematically analyzing the interactive influence relationships between various elements, combined with big data analysis technology, an interactive influence map is constructed, so that the system has a global cognitive ability of the effect of ingredient combinations. This map not only improves the scientificity and rationality of formula design, but also provides structured knowledge support for subsequent recommendations; by obtaining multi-dimensional health data of target users and combining big data analysis technology, it ensures that the recommendation system can fully perceive the user's current physical condition, lifestyle and psychological characteristics, so that the system has the ability to accurately identify individual differences of users, laying a data foundation for building formulas that are highly consistent with individual needs, and significantly improving the credibility and professionalism of health-oriented recommendations; using the collected user health data, on the one hand, mandatory ingredient sorting is implemented In addition to forming hard ingredient constraints based on the user's medical history or physical condition to avoid potential risks, on the other hand, the user's functional preferences in metabolic regulation, energy enhancement, emotional regulation, etc. are extracted to form a structured health preference portrait. The health preference portrait comprehensively considers both taboos and needs, making the recommendation both safe and controllable, and with a clear functional orientation, thereby significantly improving the effectiveness and practicality of the recommended content; combining the user's health preference portrait with the interactive influence map, performing formula adaptation analysis, and finally generating a recommended coffee recipe that meets the user's individual health needs, has a reasonable ingredient synergy logic, and a balanced flavor structure. This not only integrates the dual support of individual portraits and ingredient knowledge, but also realizes an accurate, efficient, and feasible formula generation process, significantly improving the personalization level and health intervention capabilities of the recommendation system.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flowchart of a method for recommending coffee recipes based on user health data provided in an embodiment of the present application.

[0011] Figure 2 Schematic diagram of the coffee recipe recommendation system based on user health data provided in an embodiment of the present application.

[0012] Explanation of the accompanying drawings: interactive impact analysis module 10, health data reading module 20, preference portrait construction module 30, recipe adaptation analysis module 40. DETAILED DESCRIPTION

[0013] The embodiments of the present application provide a coffee recipe recommendation method and system based on user health data, which solves the technical problem that the coffee recipe recommendation method in the prior art relies on the user's historical purchase records or taste preferences to make recommendations, and is unable to make timely adjustments when the user's health status or behavioral habits change, resulting in reduced recommendation accuracy and user satisfaction.

[0014] After introducing the basic principles of this application, various non-limiting embodiments of this application will be specifically described below in conjunction with the accompanying drawings. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0015] Example 1, as Figure 1 As shown, the embodiment of the present application provides a coffee recipe recommendation method based on user health data, the method comprising:

[0016] Collect categories of coffee blending elements, conduct interactive impact analysis on various coffee blending elements, and construct an interactive impact map.

[0017] Collect categories of coffee blending elements, including but not limited to coffee bean type, grinding fineness, roasting degree, ingredients, water temperature, brewing time, coffee concentration, etc. The different properties of each element will have different effects on the flavor and health benefits of coffee. Conduct interaction analysis on various coffee blending elements, including synergistic or antagonistic effect analysis between any two or more coffee blending elements, where a synergistic effect refers to when two or more coffee blending elements are combined, their overall effect is greater than the sum of their individual effects; an antagonistic effect refers to when two or more blending elements work together, one of the ingredients weakens or offsets the effect of another ingredient. By analyzing these interaction effects, it is possible to match coffee ingredients more scientifically, avoid invalid or conflicting combinations, and enhance beneficial functional combinations.

[0018] Based on the results of the interaction impact analysis, we use graph theory to construct an interaction impact map. Each node in the interaction impact map represents a coffee blending element, and each edge represents the interaction relationship between elements. The edge weight indicates the strength of the interaction effect. The interaction impact map shows how different elements interact with each other, helping to adapt and optimize subsequent recipes.

[0019] Read the target user's health data, including physiological parameters, behavior records, disease history data, and emotional state scores.

[0020] Relevant user health data is read from the target user's personal health file. These data may come from electronic health records, wearable devices or data entered by the user. Among them, physiological parameters such as blood pressure, blood sugar, heart rate, weight, body temperature, etc. These physiological parameters directly affect the user's health status, especially the reaction to caffeine, sugar, etc. in coffee; behavioral records include the user's eating habits, exercise records, work and rest time, etc., which can reflect the user's lifestyle and affect the way and amount of coffee intake. For example, excessive exercise or irregular sleep may change the user's tolerance to coffee; medical history data includes information on the user's chronic disease history, allergy history, cardiovascular disease, diabetes, etc. The user's health status largely determines which coffee ingredients are suitable or unsuitable for their physique, and avoids recommending ingredients that will have an adverse effect on the user's health; the emotional state score will affect the user's perception of coffee, especially in terms of psychological reactions such as excitement and anxiety caused by caffeine.

[0021] The collected data is preprocessed, including denoising, standardization, and filling in missing values, to ensure the accuracy and consistency of the data. Then, through data fusion technologies such as multimodal data fusion, physiological parameters, behavioral records, disease history, and emotional state scores are integrated into a unified user health data to provide comprehensive information for subsequent analysis.

[0022] Based on the user health data, hard component constraints and functional preference feature extraction are performed to construct a user health preference profile.

[0023] Based on user health data, hard ingredient constraints are implemented. That is, based on the user's health status, medical history, physiological parameters and other data, certain coffee ingredients or combinations that are detrimental to the user's health are excluded. For example, if the target user has a history of heart disease, they need to avoid formulas with high caffeine content; if the user has diabetes, they need to limit sugar intake. By comparing the user's health data with the restrictions on coffee ingredients, ingredients that are unacceptable or unsuitable for the user are screened out, forming hard ingredient constraints.

[0024] Based on the user's health data, functional preference features are further extracted, that is, the personalized needs of the target user. For example, if the user focuses on specific health goals such as weight loss, relevant ingredient preferences can be extracted, such as low-calorie formulas, such as sugar-free coffee. These functional demand features are extracted based on the target user's physiological, behavioral and emotional data, reflecting the target user's specific needs and preferences in diet and health management.

[0025] Through a comprehensive analysis of hard ingredient constraints and functional preference characteristics, a user health preference portrait is constructed. The user health preference portrait reflects the target user's health needs, taboo ingredients, functional preferences, etc., and will serve as the core basis for subsequent recipe recommendations to ensure that the recommended coffee recipe can meet the user's health and functional needs.

[0026] A recipe adaptation analysis is performed in combination with the interaction influence map and the user health preference profile to generate a recommended coffee recipe for the target user.

[0027] The interactive impact map is combined with the user's health preference profile to conduct a recipe adaptation analysis. Specifically, based on the user's health preference profile, coffee ingredients that meet their health needs are screened out, and taboo ingredients are excluded, such as avoiding high caffeine that is harmful to the heart and acidic ingredients that cause stomach discomfort. On the basis of meeting health requirements, functional preferences are further optimized, such as adjusting the proportion of coffee ingredients, adding or reducing certain ingredients to meet the user's energy needs, antioxidant needs, etc. According to the user's taste preferences, such as preference for rich or light taste, the flavor characteristics of the coffee are adjusted. This process can be combined with the user's historical beverage data to determine their preference for different flavors and make flavor optimization. Based on the above recipe adaptation analysis, a set of personalized recommended coffee recipes is finally generated. This recommended coffee recipe not only meets the user's health requirements, but can also be optimized according to the user's functional needs and taste preferences, making the recommendation results more accurate and personalized.

[0028] Furthermore, the interaction analysis includes the analysis of synergistic or antagonistic effects between any two or more coffee blending elements.

[0029] The synergistic effect refers to the fact that when two or more coffee blending elements are combined, their overall effect is greater than the sum of their individual effects; the antagonistic effect refers to the fact that when two or more blending elements work together, one component weakens or offsets the effect of another component. For example, the acid-base neutralization effect, that is, acidic coffee beans (such as Ethiopian Arabica) often have a strong sour taste, and certain ingredients (such as milk) can neutralize the acidity of coffee due to their alkaline properties, thereby reducing the acidity of coffee, making the taste softer and less sharp. For example, when milk, almond milk and other milks are combined with acidic coffee beans, the sour taste will be neutralized, which is suitable for consumers who do not like strong sour taste; the functional enhancement effect, that is, the ingredients in some formulas can enhance the functions of other ingredients. For example, when MCT oil (medium-chain fatty acid oil) is combined with caffeine, MCT oil can delay the release of caffeine and slow down the metabolic rate of caffeine in the body, which can not only make the effect of caffeine more lasting, but also avoid the anxiety and discomfort caused by caffeine. The addition of MCT oil can make the stimulating effect of caffeine more stable, thereby enhancing the overall metabolic promotion effect.

[0030] Furthermore, we conduct an interactive impact analysis on various coffee blending elements and construct an interactive impact map, including:

[0031] Step 1: Extract the first blending element and the second blending element from each type of coffee blending element; Step 2: Collect historical coffee blending data, where each piece of data carries a performance label; Step 3: Based on the historical coffee blending data, perform data extraction under quantitative conditions with the first blending element and the second blending element as variables and other blending elements, perform synergistic or antagonistic effect analysis based on the extracted data, and generate a first interaction effect relationship; Step 4: Based on the first interaction effect relationship, construct a first interaction influence map with the first blending element and the second blending element as nodes and the first interaction effect relationship as an edge; Step 5: Continue to extract other element combinations from the various types of coffee blending elements, execute steps 2 to 3, supplement the map based on the first interaction influence map, and generate the interaction influence map.

[0032] Two elements were extracted from various coffee blending elements for interaction effect analysis. These two elements were used as the first blending element and the second blending element. These two elements can be common coffee blending element pairs, such as coffee beans and milk, coffee beans and sugar, caffeine and MCT oil, etc.; they can also be functional combinations. For example, for a formula that enhances energy effects, caffeine and MCT oil are selected as the first and second blending elements. For a formula that improves antioxidant effects, coffee beans rich in antioxidants and matched antioxidant ingredients are selected.

[0033] A large amount of historical coffee blending data is collected from existing coffee recipe databases or experiments. This data includes the specific combinations of different blending elements, recipes, and corresponding results. Each data item carries a performance tag that describes the key effects of the coffee recipe, such as: flavor and taste, such as sweetness, sourness, richness, and smoothness; metabolic effects, such as the refreshing effect of caffeine, the metabolic effects of MCT oil, and its antioxidant effects; emotional state effects, such as the impact of coffee on anxiety, calmness, or focus; physiological reactions, such as blood sugar changes and digestive system reactions. In this way, the effects of different elements can be systematically understood, providing a sufficient reference for subsequent data extraction and synergistic effect analysis.

[0034] Based on historical coffee blending data, the data of the first blending element and the second blending element are extracted and used as variables for interaction effect analysis. For other blending elements, they are controlled as quantitative conditions. Specifically, the first blending element and the second blending element are used as independent variables, and other ingredients, such as sugar, milk, spices, etc., are used as control variables to ensure that only the interaction between the first and second blending elements is analyzed. By screening historical data that meets specific conditions (such as health effects, flavor scores, etc.), the representativeness and validity of the extracted samples are ensured.

[0035] After data extraction, an analysis of synergistic or antagonistic effects is conducted. Specifically, the first blending element is checked to see if it produces a stronger positive effect when combined with the second blending element than when used alone. For example, whether the combination of caffeine and MCT oil can provide a more sustained energy boost, rather than the simple effect of a single ingredient. The system also checks whether the two blending elements offset each other's effects. For example, certain high-sugar formulas may weaken the metabolic effects of caffeine, thereby affecting the overall functional effect. Based on the analysis results, a first interaction effect relationship is generated between the first blending element and the second blending element. The first interaction effect relationship is used to demonstrate the degree and type of interaction between different blending elements, providing data support for subsequent formula adaptation and helping the system accurately recommend the best formula combination.

[0036] Define the nodes in the interaction influence map, with each node representing a coffee blend element. These elements can include coffee bean type, milk, sugar, caffeine, spices, oils, and more. Extract the first and second blend elements from the selected blend elements and use them as the basic nodes in the map. In the map, edges represent the interactions between two blend elements. These edges are defined based on the first interaction effect relationship, each representing a synergistic or antagonistic effect between the two blend elements. The edge weight can be set based on the strength of the interaction effect. For example, if there is a synergistic effect between the two elements, the edge weight is positive, indicating that they promote each other in some way, such as enhancing flavor or health benefits. If there is an antagonistic effect between the two elements, the edge weight is negative, indicating that they may inhibit each other, such as some ingredients may offset the health effects or taste characteristics of other ingredients. Based on these nodes and edges, a first interaction influence map is constructed. Each node in the map represents a blend element, and the edges represent the interactions between them. The structure of the map clearly illustrates the interaction patterns between different elements.

[0037] After completing the interaction effect analysis between the first blending element and the second blending element, continue to extract new combinations from other types of coffee blending elements and include these new element combinations in the graph analysis together with the previous combinations. For example, in addition to the combination of coffee beans and milk, you can also analyze different element combinations such as coffee beans and spices, sugar and milk, or caffeine and MCT oil. Perform the same operations as steps two and three on the new blending element combinations. Based on the existing first interaction influence map, add the newly extracted element combinations and their interaction effects to the map. For each new pair of blending element combinations, generate new nodes and edges, and update the corresponding relationships in the map. After continuous element combination analysis and map supplementation, a complete interaction influence map is finally formed. This interaction influence map not only shows the relationship between the blending elements, but also calibrates the strength of the synergistic and antagonistic effects between each pair of elements through weights.

[0038] Furthermore, combining the interaction influence map and the user health preference profile to perform recipe adaptation analysis and generate a recommended coffee recipe for the target user, including:

[0039] Collect historical recipe data to build a recipe database; locate recipe taboos based on the user's health preference profile; delete taboo recipes from the recipe database using the recipe taboos to generate a recipe retention database; optimize health adaptation, function adaptation, and flavor adaptation for each recipe in the recipe retention database based on the user's health preference profile to generate the recommended coffee recipe.

[0040] Collect a large amount of historical recipe data. Each recipe includes recipe composition, flavor labels, health labels, user feedback, etc. These historical recipe data can come from a variety of sources, such as user feedback and ratings, experimental data, professional barista and coffee culture databases, etc. The collected historical recipe data is standardized and structured, and then stored in the recipe database to ensure that the data can be quickly retrieved and updated.

[0041] Based on the user health preference profile in the previous step, analyze the target user's health status, taboo ingredients, functional requirements and other information to locate all possible formula taboos. These formula taboos involve ingredients in the formula, such as excessive caffeine content, excessive sugar, etc., or certain combinations of formulas, such as the interaction between caffeine and specific drugs.

[0042] The recipe database is deleted based on recipe contraindications. This means that all recipes containing contraindicated ingredients or those that are unsuitable for the user are removed from the recipe database. For example, if a user cannot consume certain ingredients, such as sugar or milk, recipes containing these ingredients are deleted. If a user has high blood pressure and is contraindicated, all high-caffeine recipes are removed. After deleting contraindicated recipes, the remaining recipes that meet the user's health requirements, taste preferences, and functional needs constitute the recipe retention database. This database only contains recipes suitable for the target user, ensuring that each user receives personalized, healthy, and flavor-matched coffee recommendations.

[0043] After deleting the taboo formulas, continue to optimize the formulas according to the user's health needs. For example, if the user needs to lose weight or control blood sugar, recommend low-calorie, low-sugar formulas; if the user has antioxidant needs, give priority to formulas rich in antioxidant ingredients (such as green coffee beans, cinnamon, etc.). After health adaptation, the remaining formulas are further optimized according to the user's functional needs, such as refreshing, enhancing metabolism, promoting digestion, and increasing energy. For example, if the user needs to enhance metabolism, give priority to formulas containing MCT oil and caffeine, which can delay the release of caffeine and improve the metabolic effect. After the health and functional adaptation is completed, flavor adaptation is finally performed. According to the user's flavor preferences, such as strong or light taste, sour or sweet taste, etc., the formula is adjusted. For example, if the user prefers strong coffee, dark roasted coffee beans can be recommended; if the user likes a light taste, lightly roasted beans can be recommended or the addition of appropriate amounts of milk, cream and other ingredients can be recommended. Finally, based on the optimization of health adaptation, functional adaptation and flavor adaptation, a set of recommended coffee recipes that best meet user needs are generated. These recommended coffee recipes not only meet the user's health requirements and improve functional effects, but also provide a drinking experience that meets the user's flavor preferences.

[0044] Furthermore, in the process of optimizing health adaptation, function adaptation and flavor adaptation for each formula, the adaptation priority of health adaptation is higher than that of function adaptation, which is higher than that of flavor adaptation.

[0045] During the recipe optimization process, health compatibility is prioritized, meaning any recipe recommendation must first meet the user's health needs and contraindications. Once a recipe passes health compatibility, functional adaptation is then considered based on the user's personalized health goals, such as energy boosting, weight loss, and antioxidant protection. Following health and functional adaptation, flavor adaptation is finalized to ensure the recommended recipe meets the user's taste preferences. By optimizing flavor adaptation, we ultimately ensure that users not only have their health and functional needs met, but also enjoy a pleasant taste experience.

[0046] Furthermore, based on the user's health preference profile, each recipe in the recipe retention database is optimized for health adaptation, function adaptation, and flavor adaptation to generate the recommended coffee recipe, including:

[0047] Each recipe in the recipe retention database is sorted by health adaptation and functional adaptation scores based on the user's health preference profile, and the historical behavior data of the target user is simultaneously introduced to sort by flavor adaptation scores to generate a health adaptation sequence, a functional adaptation sequence, and a flavor adaptation sequence; the intersection recipes of the health adaptation sequence, the functional adaptation sequence, and the flavor adaptation sequence that meet the preset adaptation indicators and adaptation priorities are extracted to generate the recommended coffee recipe.

[0048] Based on the user's health preference profile, a health adaptation score is given to each recipe in the recipe retention database. Specifically, the ingredients in each recipe are checked for health adaptation to ensure that the recipe does not contain ingredients that are taboo for the user or do not meet health needs, such as high sugar, high caffeine, etc. Each recipe is scored according to the user's health goals, such as anti-oxidation, weight loss, energy increase, etc. Recipes with high satisfaction will receive higher health adaptation scores. According to the health adaptation score, all recipes are sorted by health adaptation to generate a health adaptation sequence, in which recipes with high health adaptation are ranked at the front.

[0049] After health adaptation, each formula is scored for functional adaptation. Specifically, the functional ingredients in the formula are scored based on the user's functional needs, such as improving metabolism and promoting digestion. For example, if the user needs to improve metabolism, a formula containing MCT oil and caffeine will score higher. Each formula is comprehensively evaluated to see how well it meets the user's functional needs. Formulas with higher scores are recommended first. According to the functional adaptation score, the formulas are sorted to generate a functional adaptation sequence, with formulas with higher functional adaptation scores ranked first.

[0050] The target user's historical behavior data is introduced to perform flavor adaptation score sorting. Specifically, the target user's historical behavior data is analyzed, including the target user's previous beverage selection, ratings, comments, etc., to understand their flavor preferences, such as sour, sweet, and strong taste. Based on the user's preferred flavor, the flavor matching degree of each recipe is calculated. If the flavor of the recipe meets the user's historical preferences, a higher flavor adaptation score is obtained. According to the flavor adaptation score, the recipes are sorted to generate a flavor adaptation sequence, and the recipes with high flavor adaptation scores are ranked first.

[0051] Intersection recipes are extracted from the health, function, and flavor adaptation sequences. Intersection recipes are those that simultaneously meet all adaptation requirements. Specifically, the recipes meet the user's health requirements, avoid contraindications, and meet health goals; the recipes satisfy the user's functional needs, such as energy boosting and metabolic enhancement; and the flavor profile matches the user's flavor preferences. Intersection recipes are evaluated based on pre-set adaptation metrics. For example, health adaptation must exceed a certain threshold, such as 90%, functional adaptation must meet at least 80% of the user's functional needs, and flavor adaptation must meet at least 90% of the user's flavor preferences. Only intersection recipes that meet these pre-set adaptation metrics are retained. The retained intersection recipes are sorted based on their health, function, and flavor adaptation priorities, with health first, function second, and flavor last. Finally, after extracting the intersection recipes and sorting them by adaptation priority, a recommended coffee recipe is generated that meets all criteria. The recommended coffee recipe not only meets the user's health, function, and flavor preferences, but also provides a personalized drinking experience.

[0052] Furthermore, generating the recommended coffee recipe further includes:

[0053] If there is no intersection formula that meets the preset adaptation index among the health adaptation sequence, function adaptation sequence and flavor adaptation sequence, a first intersection formula of the health adaptation sequence and function adaptation sequence is extracted according to the adaptation priority; the historical behavior data of the target user is introduced to determine the flavor influencing factor, and an analysis of the mutual influence of flavor, health and function is performed. The flavor adjustment of the first intersection formula is performed according to the mutual influence relationship to generate the recommended coffee recipe.

[0054] If there is no intersection formula that meets the preset adaptation index in the health adaptation sequence, functional adaptation sequence and flavor adaptation sequence, for example, the health adaptation is high but the flavor is extremely mismatched, then the formula extraction is performed according to the set adaptation priority, that is, health adaptation takes priority, functional adaptation takes second place, and flavor adaptation takes last. At this time, the flavor adaptation dimension is excluded, and the intersection is only sought in the health and function dimensions. A set of first intersection formulas that meet the health and function adaptation requirements but may have differences in flavor are obtained, which serves as the basis for the next step of flavor adjustment.

[0055] Flavor influencing factors are determined based on the historical behavioral data of target users. For example, users prefer nutty flavors, accept mild sour tastes, and dislike excessive bitterness. The relationship between flavor and health status (such as digestion) and functional needs (such as improving alertness) is analyzed. For example, certain flavor ingredients may have positive or negative effects on health, such as vanilla flavor additives may contain too much sugar; certain sources of bitterness (such as caffeine concentration) enhance function, but users are sensitive. For the first intersection formula screened out, flavor adjustment is performed based on the above-mentioned mutual influence relationship. For example, the flavor compensation method is used to retain the core health and functional ingredients, and the overall flavor adaptability is improved by replacing flavor substances, such as adding spices (cinnamon, cardamom) and flavor blending agents (oat milk, etc.); the fine-tuning dosage method is used to fine-tune the ratio of bitter, sour, and sweet ingredients in the formula without destroying the function; the flavor simulation method is used to introduce low-calorie simulated flavor substances to enhance the experience without affecting health. The flavor-adjusted recipe is used as the recommended coffee recipe. This mechanism ensures that even when the three adaptations are difficult to intersect, it can flexibly adapt to the user's subjective flavor preferences based on user health and functional safety, and ultimately output a high-satisfaction recipe.

[0056] Furthermore, based on the user health data, hard component constraints and function preference feature extraction are performed to construct a user health preference profile, including:

[0057] According to the disease history data and physiological parameters in the user's health data, forced conflict matching of coffee ingredients is performed based on big data to generate hard ingredient constraints; functional preference analysis is performed based on the physiological parameters, behavior records and emotional state scores in the user's health data to obtain functional preference characteristics; the user's health preference portrait is generated based on the hard ingredient constraints and the functional preference characteristics.

[0058] Medical history data includes data on hypertension, diabetes, and stomach problems, and physiological parameters include blood pressure, blood sugar, and liver function indicators. The target user's medical history data and physiological parameters are used to identify coffee blending elements that have potential adverse reactions or health risks. For example, a coffee ingredient and disease risk matching model is constructed based on big data. The model is derived from medical literature, nutrition research data, and historical user feedback. With the support of this model, the user's health information is compared one by one with each ingredient in the coffee formula to identify ingredients that are potentially harmful to the user's health, such as the stimulating effect of high caffeine ingredients on people with hypertension, and the stimulating effect of acidic ingredients on stomach diseases. These ingredients are then listed as strictly prohibited or restricted ingredients to form a clear ingredient blacklist, that is, a set of ingredients that the target user must exclude or control the dosage in subsequent recommended formulas as a hard ingredient constraint.

[0059] Physiological parameters include metabolic rate, body fat percentage, and blood sugar levels, which can reflect whether the user suffers from slow metabolism or low energy levels. Behavioral records include exercise frequency, daily routines, and dietary habits, which can reflect whether the user is sedentary, stays up late, or overworked. Emotional state scores quantify psychological manifestations such as anxiety, irritability, and mental fatigue, which can reflect the need for support from mood-regulating formulas. Through comprehensive analysis of this data, combined with the knowledge of the efficacy of existing functional ingredients, such as MCT oil for enhanced metabolism and L-theanine for mood stabilization, the target user's current functional preference characteristics are established, including: metabolic enhancement needs (suitable for those who are sedentary, have low metabolic rates, and are trying to control their weight); emotional regulation needs (suitable for those who are anxious, stressed, and have mood swings); and energy enhancement needs (suitable for those who engage in mental work and are prone to fatigue).

[0060] The hard ingredient constraints and functional preference characteristics are integrated to form a complete and structured user health preference portrait. The user health preference portrait not only includes specific substances that the target users should not consume, but also clearly points out the functional experience they hope to obtain through coffee, such as promoting metabolism and stabilizing mood, thereby constituting a dual guidance for personalized formula matching.

[0061] Furthermore, after generating the recommended coffee recipe, the process further includes:

[0062] According to a preset statement structure, an explanatory recommendation statement is constructed using the user health data, the hard component constraints, and the functional preference characteristics to generate a target explanatory statement; and the target explanatory statement is fed back to the target user together with the recommended coffee recipe.

[0063] Several common preset sentence structures are set. Variable fields are reserved in the preset sentence structures for filling in user health data, hard component constraints, and functional preference features. The above fields are filled in the preset sentence structure through rule matching or natural language generation models, and the sentence fluency, information coherence, and professional expression are automatically optimized according to semantic logic to form a target explanation sentence. For example, for a user who suffers from excessive gastric acid and hopes to improve fatigue, the following target explanation sentence may be generated: Since you have a problem with high gastric acid secretion, we have ruled out the highly irritating strong caffeine. At the same time, considering your need to refresh yourself and fight fatigue, we recommend adding low-caffeine coffee and coenzyme Q10 to the formula to achieve a mild refreshing and stomach-protecting effect.

[0064] The target explanation statement is bound to the recommended coffee recipe and fed back to the target user. By presenting the recommended content and the reason for the recommendation simultaneously, the target user not only knows what to drink, but also understands why to drink it. This design can effectively improve user acceptance and stickiness, while providing semantic support for subsequent user feedback collection, behavior record updates and model iterations, thus forming a closed-loop optimization.

[0065] In summary, the coffee recipe recommendation method based on user health data provided by the embodiments of the present application has the following technical effects:

[0066] By comprehensively collecting the categories of coffee blending elements and systematically analyzing the interactive influence relationships between various elements, combined with big data analysis technology, an interactive influence map is constructed, so that the system has a global cognitive ability of the effect of ingredient combinations. This map not only improves the scientificity and rationality of formula design, but also provides structured knowledge support for subsequent recommendations; by obtaining multi-dimensional health data of target users and combining big data analysis technology, it ensures that the recommendation system can fully perceive the user's current physical condition, lifestyle and psychological characteristics, so that the system has the ability to accurately identify individual differences of users, laying a data foundation for building formulas that are highly consistent with individual needs, and significantly improving the credibility and professionalism of health-oriented recommendations; using the collected user health data, on the one hand, mandatory ingredient sorting is implemented In addition to forming hard ingredient constraints based on the user's medical history or physical condition to avoid potential risks, on the other hand, the user's functional preferences in metabolic regulation, energy enhancement, emotional regulation, etc. are extracted to form a structured health preference portrait. The health preference portrait comprehensively considers both taboos and needs, making the recommendation both safe and controllable, and with a clear functional orientation, thereby significantly improving the effectiveness and practicality of the recommended content; combining the user's health preference portrait with the interactive influence map, performing formula adaptation analysis, and finally generating a recommended coffee recipe that meets the user's individual health needs, has a reasonable ingredient synergy logic, and a balanced flavor structure. This not only integrates the dual support of individual portraits and ingredient knowledge, but also realizes an accurate, efficient, and feasible formula generation process, significantly improving the personalization level and health intervention capabilities of the recommendation system.

[0067] Example 2 is based on the same inventive concept as the coffee recipe recommendation method based on user health data in the previous embodiment. Figure 2 As shown, an embodiment of the present application provides a coffee recipe recommendation system based on user health data, the system comprising:

[0068] Interaction impact analysis module 10, used to collect coffee blending element categories, perform interaction impact analysis on various coffee blending elements, and construct an interaction impact map;

[0069] The health data reading module 20 is used to read the user health data of the target user, including physiological parameters, behavior records, disease history data and emotional state scores;

[0070] a preference profile building module 30 for performing hard component constraints and extracting functional preference features based on the user health data to build a user health preference profile;

[0071] The recipe adaptation analysis module 40 is used to perform recipe adaptation analysis based on the interaction influence map and the user health preference profile to generate a recommended coffee recipe for the target user.

[0072] Furthermore, the interaction analysis includes the analysis of synergistic or antagonistic effects between any two or more coffee blending elements.

[0073] Furthermore, the interaction impact analysis module 10 is configured to perform the following steps:

[0074] Step 1: Extract the first blending element and the second blending element from each type of coffee blending element; Step 2: Collect historical coffee blending data, where each piece of data carries a performance label; Step 3: Based on the historical coffee blending data, perform data extraction under quantitative conditions with the first blending element and the second blending element as variables and other blending elements, perform synergistic or antagonistic effect analysis based on the extracted data, and generate a first interaction effect relationship; Step 4: Based on the first interaction effect relationship, construct a first interaction influence map with the first blending element and the second blending element as nodes and the first interaction effect relationship as an edge; Step 5: Continue to extract other element combinations from the various types of coffee blending elements, execute steps 2 to 3, supplement the map based on the first interaction influence map, and generate the interaction influence map.

[0075] Furthermore, the recipe adaptation analysis module 40 is configured to perform the following steps:

[0076] Collect historical recipe data to build a recipe database; locate recipe taboos based on the user's health preference profile; delete taboo recipes from the recipe database using the recipe taboos to generate a recipe retention database; optimize health adaptation, function adaptation, and flavor adaptation for each recipe in the recipe retention database based on the user's health preference profile to generate the recommended coffee recipe.

[0077] Furthermore, in the process of optimizing health adaptation, function adaptation and flavor adaptation for each formula, the adaptation priority of health adaptation is higher than that of function adaptation, which is higher than that of flavor adaptation.

[0078] Furthermore, the recipe adaptation analysis module 40 is configured to perform the following steps:

[0079] Each recipe in the recipe retention database is sorted by health adaptation and functional adaptation scores based on the user's health preference profile, and the historical behavior data of the target user is simultaneously introduced to sort by flavor adaptation scores to generate a health adaptation sequence, a functional adaptation sequence, and a flavor adaptation sequence; the intersection recipes of the health adaptation sequence, the functional adaptation sequence, and the flavor adaptation sequence that meet the preset adaptation indicators and adaptation priorities are extracted to generate the recommended coffee recipe.

[0080] Furthermore, the recipe adaptation analysis module 40 is configured to perform the following steps:

[0081] If there is no intersection formula that meets the preset adaptation index among the health adaptation sequence, function adaptation sequence and flavor adaptation sequence, a first intersection formula of the health adaptation sequence and function adaptation sequence is extracted according to the adaptation priority; the historical behavior data of the target user is introduced to determine the flavor influencing factor, and an analysis of the mutual influence of flavor, health and function is performed. The flavor adjustment of the first intersection formula is performed according to the mutual influence relationship to generate the recommended coffee recipe.

[0082] Furthermore, the preference profile building module 30 is configured to perform the following steps:

[0083] According to the disease history data and physiological parameters in the user's health data, forced conflict matching of coffee ingredients is performed based on big data to generate hard ingredient constraints; functional preference analysis is performed based on the physiological parameters, behavior records and emotional state scores in the user's health data to obtain functional preference characteristics; the user's health preference portrait is generated based on the hard ingredient constraints and the functional preference characteristics.

[0084] Furthermore, the recipe adaptation analysis module 40 is configured to perform the following steps:

[0085] According to a preset statement structure, an explanatory recommendation statement is constructed using the user health data, the hard component constraints, and the functional preference characteristics to generate a target explanatory statement; and the target explanatory statement is fed back to the target user together with the recommended coffee recipe.

[0086] Through the above detailed description of the coffee recipe recommendation method based on user health data in this specification, those skilled in the art can clearly understand the coffee recipe recommendation system based on user health data in this embodiment. Since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0087] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A coffee recipe recommendation method based on user health data, characterized in that: The method comprises: Collect coffee blending element categories, conduct interactive impact analysis on various coffee blending elements, and construct an interactive impact map; Read the target user's health data, including physiological parameters, behavior records, medical history data, and emotional state scores; Based on the user health data, hard component constraints and functional preference feature extraction are performed to build a user health preference profile; A recipe adaptation analysis is performed in combination with the interaction influence map and the user health preference profile to generate a recommended coffee recipe for the target user.

2. The coffee recipe recommendation method based on user health data according to claim 1, characterized in that: Interaction analysis includes the analysis of synergistic or antagonistic effects between any two or more coffee blending elements.

3. The coffee recipe recommendation method based on user health data according to claim 2, characterized in that: Conduct interactive impact analysis on various coffee blending elements and construct interactive impact maps, including: Step 1: Extracting the first blending element and the second blending element from various coffee blending elements; Step 2: Collect historical coffee blending data. Each piece of data in the historical coffee blending data carries a performance label. Step 3: Based on the historical coffee blending data, extract data with the first blending element and the second blending element as variables and the other blending elements as quantitative conditions, perform synergistic or antagonistic effect analysis based on the extracted data, and generate a first interactive effect relationship; Step 4: Based on the first interaction effect relationship, a first interaction influence graph is constructed with the first allocation element and the second allocation element as nodes and the first interaction effect relationship as an edge; Step 5: Continue to extract other element combinations from the various coffee blending elements, execute steps 2 to 3, and supplement the first interaction influence map to generate the interaction influence map.

4. The coffee recipe recommendation method based on user health data according to claim 1, characterized in that: Combining the interaction influence map and the user health preference profile to perform recipe adaptation analysis and generate a recommended coffee recipe for the target user, including: Collect historical recipe data to build a recipe database; Locating recipe contraindications based on the user's health preference profile; Deleting taboo recipes from the recipe database using the recipe taboo items to generate a recipe retention database; Based on the user's health preference profile, each recipe in the recipe retention database is optimized for health adaptation, function adaptation and flavor adaptation to generate the recommended coffee recipe.

5. The coffee recipe recommendation method based on user health data according to claim 4, characterized in that: In the process of optimizing health adaptation, function adaptation and flavor adaptation for each formula, the adaptation priority of health adaptation is higher than that of function adaptation, which is higher than that of flavor adaptation.

6. The coffee recipe recommendation method based on user health data according to claim 5, characterized in that: Based on the user's health preference profile, each recipe in the recipe retention database is optimized for health adaptation, function adaptation, and flavor adaptation to generate the recommended coffee recipe, including: Perform health adaptation and function adaptation score sorting based on the user's health preference profile on each recipe in the recipe retention database, and simultaneously introduce the historical behavior data of the target user to perform flavor adaptation score sorting to generate a health adaptation sequence, a function adaptation sequence, and a flavor adaptation sequence; An intersection formula that satisfies preset adaptation indicators and adaptation priorities is extracted from the health adaptation sequence, the function adaptation sequence, and the flavor adaptation sequence to generate the recommended coffee formula.

7. The coffee recipe recommendation method based on user health data according to claim 6, characterized in that: Generating the recommended coffee recipe further includes: If there is no intersection formula that meets the preset adaptation index among the health adaptation sequence, the function adaptation sequence, and the flavor adaptation sequence, extracting a first intersection formula of the health adaptation sequence and the function adaptation sequence according to the adaptation priority; The historical behavior data of the target user is introduced to determine the flavor influencing factor, and an analysis of the mutual influence of flavor, health, and function is performed. The flavor of the first intersection formula is adjusted according to the mutual influence relationship to generate the recommended coffee formula.

8. The coffee recipe recommendation method based on user health data according to claim 1, characterized in that: Based on the user health data, hard component constraints and function preference feature extraction are performed to build a user health preference profile, including: performing forced conflict matching of coffee ingredients based on big data according to the disease history data and physiological parameters in the user's health data, and generating hard ingredient constraints; Performing function preference analysis based on physiological parameters, behavior records, and emotional state scores in the user's health data to obtain function preference characteristics; The user health preference profile is generated based on the hard component constraints and the functional preference characteristics.

9. The coffee recipe recommendation method based on user health data according to claim 1, characterized in that: After generating the recommended coffee recipe, the method further includes: According to a preset sentence structure, constructing an explanatory recommendation sentence based on the user health data, the hard component constraints, and the function preference characteristics to generate a target explanatory sentence; The target explanation sentence and the recommended coffee recipe are fed back to the target user.

10. A coffee recipe recommendation system based on user health data, characterized in that: A system for implementing the coffee recipe recommendation method based on user health data according to any one of claims 1 to 9, comprising: The interactive impact analysis module is used to collect coffee blending element categories, conduct interactive impact analysis on various coffee blending elements, and construct an interactive impact map; The health data reading module is used to read the user health data of the target user, including physiological parameters, behavior records, disease history data and emotional state scores; A preference profile building module, configured to perform hard component constraints and extract functional preference features based on the user health data to build a user health preference profile; The recipe adaptation analysis module is used to perform recipe adaptation analysis based on the interaction influence map and the user health preference profile to generate a recommended coffee recipe for the target user.

Images