Meal recommendation method and system based on big data analysis

Abstract

The invention provides a food recommendation method and system based on big data analysis. The food recommendation method based on big data analysis comprises the steps of obtaining identity information of a target user, judging whether the target user is a new user or not, and if yes, recommending food to the target user according to a food recommendation set obtained by analyzing first historical consumption data of a restaurant; otherwise, acquiring second historical consumption data of the target user in the restaurant, determining a target recommendation list according to the second historical consumption data, and recommending food to the target user according to the target recommendation list. By analyzing the historical consumption data of the restaurant, the food recommendation set suitable for food recommendation to people with different taste preferences is determined, and compared with recommendation objects obtained through artificial experience analysis, the food recommendation set obtained through big data technology analysis has higher pertinence and representativeness.

Description

Technical Field

[0001] The present invention relates to the technical field of meal recommendation, and in particular to a meal recommendation method and system based on big data analysis. Background Art

[0002] In the catering industry, the quality of service is one of the key factors that affect the popularity of a restaurant. When ordering food in a restaurant, some users may find it difficult to make a decision when faced with a wide range of food. Generally, you can get relevant suggestions by asking the restaurant's service staff. However, the restaurant's service staff recommends some popular food to users based on their own experience. The recommended food can help users who have difficulty ordering food to a certain extent. In fact, the data analyzed by the service staff based on their own experience has certain limitations. Different customers have different preferences for different flavors of food. Therefore, this recommendation scheme based on artificial experience is not comprehensive enough. It is difficult to effectively use the restaurant's massive consumption data to obtain a more personalized recommendation scheme. The efficiency of recommending food to users is not high and the pertinence is not strong. Summary of the invention

[0003] The present application provides a meal recommendation method and system based on big data analysis, aiming to solve at least one technical problem existing in the above-mentioned background technology.

[0004] On the one hand, the present application provides a method for recommending food based on big data analysis, including: obtaining identity information of a target user, determining whether the target user is a new user, and if so, recommending food to the target user based on a food recommendation set obtained by analyzing first historical consumption data of a restaurant; otherwise, obtaining second historical consumption data of the target user in the restaurant, determining a target recommendation list based on the second historical consumption data, and recommending food to the target user based on the target recommendation list; The recommended set of dishes for a restaurant is determined in the following way: Get the restaurant's flavor label information and each dish's food label information; According to the first historical consumption data of the restaurant, the consumption times of each dish are counted respectively, and the dishes are sorted according to the consumption times to generate a total recommendation list; According to the taste label information and the food label information, a food collection with each taste label as the main body is established; For any set of dishes, all dishes in the set are sorted based on the number of times they are consumed, and a recommended sublist corresponding to the set of dishes is obtained; Traverse all meal sets, obtain the recommended sublist corresponding to each meal set, and construct a meal recommendation set based on the total recommendation list and multiple recommended sublists.

[0005] Furthermore, making meal recommendations to target users based on the meal recommendation set includes: Obtaining user preference information, and filtering a first reference list and / or a second reference list from the meal recommendation set according to the preference information; The dishes recorded in the first reference list in the recommended total list are removed to obtain a third reference list, and the dishes recorded in the second reference list are reordered according to the recommended total list to obtain a fourth reference list; Make meal recommendations to the target user based on the third reference list and / or the fourth reference list.

[0006] Further, determining a target recommendation list according to the second historical consumption data includes: Extracting the food information included in each order in the second historical consumption data, determining the food label information of each order, and respectively counting the type of food labels included in each order and the first frequency proportion of each food label, as well as the taste labels corresponding to the food included in all orders and the second frequency proportion of each taste label; The priority of each food label is calculated using the following formula: , where is the priority of the food label, For food labels The first frequency percentage in orders, For frequency The normalization parameter of is the quantity of the order; After calculating the priority of each meal label, all meal labels included in the second historical consumption data are sorted according to the priority of each meal label to obtain a label list; Determine the weight parameter for each food label in the label list , for any food label , is the number of items in the food label that meet the condition , is the total number of food labels in the label list; According to the weight parameter of each food label in the label list , the recommended value of each dish in the restaurant is determined by the following formula: , where Indicates the recommended value. is the total number of food labels for the food, For the first Food labels, For the first The weight parameter of each food label, The second frequency ratio of the food’s taste label; Sort the dishes in the restaurant according to the recommendation value of each dish to obtain a target recommendation list.

[0007] Furthermore, before making meal recommendations to the target user according to the target recommendation list, the method also includes: determining the total number of meals consumed by the target user in the restaurant according to the second historical consumption data of the target user, and judging whether the total number of meals consumed is greater than a preset evaluation threshold; if so, making meal recommendations to the target user according to the target recommendation list; otherwise, making meal recommendations to the target user according to the target recommendation list and the meal recommendation set.

[0008] Furthermore, meal recommendations are made to the target user based on the target recommendation list and the meal recommendation set, including: filtering out a recommendation sub-list corresponding to each taste tag from the meal recommendation set based on taste tags corresponding to the meals included in all orders of the target user, sorting the meals included in one or more filtered recommendation sub-lists according to the total recommendation list to obtain an auxiliary recommendation list, and collaboratively recommending the meals recorded in the target recommendation list and the auxiliary recommendation list to the target user.

[0009] Furthermore, collaboratively recommending food to the target user includes: while recommending a food in the target recommendation list to the target user, recommending a food in the auxiliary recommendation list to the target user.

[0010] On the other hand, the present application provides a meal recommendation system based on big data analysis, and the system is applied to any of the above-mentioned meal recommendation methods based on big data analysis, including: An identity analysis module is used to obtain the identity information of the target user and determine whether the user is a new user based on the identity information; A data acquisition module, used to acquire the first historical consumption data of the restaurant and the second historical consumption data of the target user in the restaurant; A data analysis module, used to analyze the first historical consumption data of the restaurant to obtain a set of recommended dishes, and used to analyze the second historical consumption data of the target user in the restaurant to obtain a target recommendation list; The meal recommendation module is used to make meal recommendations to users based on the meal recommendation set and / or the target recommendation list.

[0011] Furthermore, the data analysis module also includes: determining the total number of meals consumed by the target user in the restaurant based on the target user's second historical consumption data in the restaurant, and generating an auxiliary recommendation list by analyzing the target user's second historical consumption data in the restaurant and a set of meal recommendations.

[0012] The present invention has at least the following technical effects: 1. The present invention determines a set of meal recommendations suitable for recommending meals to people with different taste preferences by analyzing the restaurant's historical consumption data. Compared with the recommendation objects obtained by manual experience analysis, the meal recommendation set obtained by the present invention through big data technology analysis is more targeted and representative.

[0013] 2. The present invention generates a personalized target recommendation list by analyzing the user's historical consumption data in the restaurant, recommends food to the target user, and takes into account the user's total consumption in the restaurant. When the target recommendation list is not representative due to insufficient sample data, an auxiliary recommendation list is generated based on the user's consumption data to improve the efficiency of food recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0015] Figure 1 Schematic diagram of the analysis process of the first historical consumption data in an embodiment of the present invention.

[0016] Figure 2 Schematic diagram of the structure of a meal recommendation system based on big data analysis in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical scheme and advantages of the present application clearer, some embodiments of the present application are further described in detail below in conjunction with the accompanying drawings and Examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. However, it will be appreciated by those skilled in the art that in the various embodiments of the present application, many technical details are proposed in order to enable the reader to better understand the present application. However, even without these technical details and various changes and modifications based on the following embodiments, the technical scheme claimed in the present application can also be implemented.

[0018] Embodiment 1: Embodiment 1 of the present invention provides a method for recommending food based on big data analysis, including: obtaining identity information of a target user, determining whether the target user is a new user, and if so, recommending food to the target user based on a food recommendation set obtained by analyzing first historical consumption data of a restaurant; It is worth mentioning that when a user goes to a restaurant to consume, for a new user, since he has never consumed in the restaurant, he has no relevant consumption data. In this case, the restaurant's first historical consumption data can be collected through big data technology, and by analyzing the massive data, it is possible to obtain dishes that are suitable for most people, thereby constructing a set of recommended dishes for recommending dishes to new users.

[0019] Otherwise, obtain the second historical consumption data of the target user in the restaurant, determine a target recommendation list according to the second historical consumption data, and recommend food to the target user according to the target recommendation list.

[0020] It is worth noting that for users who have already consumed in the restaurant, the user's second historical consumption data can be analyzed to obtain a personalized target recommendation list and make targeted food recommendations to the user.

[0021] The overall process of meal recommendation is roughly as described above. The following is a detailed description of the recommendation process in combination with a more detailed implementation plan.

[0022] As one of the implementation schemes, Figure 1 FIG. 1 is a schematic diagram of the analysis process of the first historical consumption data in the embodiment of the present application, see Figure 1 ,This application determines the recommended set of dishes for a restaurant through the following steps: S11, obtaining the restaurant's taste label information and the dish label information of each dish; Among them, the taste label information specifically refers to the taste of each dish in the restaurant, such as sour, spicy, sweet, and numb. The meal label information specifically refers to labels that represent the characteristic information of the meal, such as labels related to the ingredients, condiments, and cooking processes of the meal. It can be imagined that the taste label is also a type of meal label.

[0023] S12, counting the number of times each meal is consumed according to the first historical consumption data of the restaurant, sorting the meals according to the number of times they are consumed, and generating a total recommendation list; Among them, by analyzing the restaurant's massive historical data, counting the total number of times each dish is consumed, obtaining sales information for each dish, and sorting the dishes according to sales, it can, to a certain extent, represent the public's preferences for the dishes, thereby obtaining a total recommended list.

[0024] S13, establishing a meal set with each flavor label as the main body according to the flavor label information and the meal label information; Each dish corresponds to a flavor, and the dishes can be classified according to their flavors to obtain multiple dish sets with each flavor label as the main body.

[0025] S14. For any meal set, all meals in the meal set are sorted based on the number of consumption times to obtain a recommended sub-list corresponding to the meal set.

[0026] S15. Traverse all meal sets to obtain a recommendation sub-list corresponding to each meal set, and construct a meal recommendation set according to the total recommendation list and multiple recommendation sub-lists.

[0027] Specifically, the total recommendation list obtained by simply analyzing the total sales data of each meal is not very targeted when recommending meals to some users who have more complex taste requirements for meals. By combining the recommendation sub-lists corresponding to each flavor obtained by separate analysis, a meal recommendation set is constructed, and the meal recommendation set is used to make recommendations to new users of the restaurant, so that the recommendations can be more targeted.

[0028] As one implementation scheme, recommending food to a target user based on the food recommendation set specifically includes: obtaining preference information of the user, and filtering a first reference list and / or a second reference list from the food recommendation set based on the preference information; Among them, the user's preference information can be one or more favorite flavors, or one or more disliked flavors. The first reference list specifically is the statistical information of the dishes corresponding to the flavors that the user dislikes, and the second reference list specifically is the statistical information of the dishes corresponding to the flavors that the user likes.

[0029] Remove the dishes recorded in the first reference list in the total recommendation list to obtain a third reference list, reorder the dishes recorded in the second reference list according to the total recommendation list to obtain a fourth reference list, and recommend dishes to target users based on the third reference list and / or the fourth reference list.

[0030] It is worth mentioning that for new users, before making meal recommendations to them, targeted recommendations can also be made based on the user's preference information, including referring to the user's favorite flavors and dietary taboos, discarding certain meals, and making targeted recommendations to the user based on the remaining meals, thereby improving the accuracy of the recommendations.

[0031] As one implementation scheme, determining a target recommendation list according to the second historical consumption data includes: Extracting the food information included in each order in the second historical consumption data, determining the food label information of each order, and respectively counting the type of food labels included in each order and the first frequency proportion of each food label, as well as the taste labels corresponding to the food included in all orders and the second frequency proportion of each taste label; The priority of each food label is calculated using the following formula: , where is the priority of the food label, For food labels The first frequency percentage in orders, For frequency The normalization parameter of is the quantity of the order; It is worth noting that simply analyzing the total number of food labels in all orders has certain limitations. For two food labels with the same total number, if one of the food labels appears frequently in a few orders, while the other food label is discretely distributed in multiple orders with a lower frequency, the user's preferences for the two food labels may be very different in these two cases. The high-frequency appearance of food labels may indicate that the user has chosen more food with the label in some special cases. For example, when the user is eating with friends, the friend likes to eat sweet food, so the sweet label appears frequently in a small number of orders. Another label, such as spicy, although the frequency of appearance in each order is not high, there are multiple orders with this label. In this case, it is more appropriate to recommend food with spicy-related labels to users than to recommend food with sweet-related labels to users. Therefore, the above formula comprehensively analyzes the frequency of occurrence of each food label in each order, as well as the number of orders containing the food label, and calculates the priority of the food label, which is highly representative.

[0032] After calculating the priority of each meal label, all meal labels included in the second historical consumption data are sorted according to the priority of each meal label to obtain a label list; Determine the weight parameter for each food label in the label list , for any food label , is the number of items in the food label that meet the condition , is the total number of food labels in the label list; It is worth noting that the weight parameter of each food label can be a set of normalized parameters, that is, , and the higher the priority of the food label, the larger the corresponding weight parameter.

[0033] According to the weight parameter of each food label in the label list , the recommended value of each dish in the restaurant is determined by the following formula: , where Indicates the recommended value. is the total number of food labels for the food, For the first Food labels, For the first The weight parameter of each food label, The second frequency ratio of the food’s taste label; Sort the dishes in the restaurant according to the recommendation value of each dish to obtain a target recommendation list.

[0034] It is worth noting that in the process of measuring the recommendation degree of dishes by numerical values, the dish labels or taste labels corresponding to each dish are considered separately, and the final numerically obtained recommendation values ​​have certain limitations, because each flavor of dishes corresponds to multiple dish labels. Even if the taste is the same, it does not mean that the user will definitely like a dish if the taste alone meets the requirements. Similarly, when multiple dish labels of a dish meet the user's intention, if they do not meet the user's taste, the user is still unlikely to choose the dish. Therefore, the above formula is used to comprehensively analyze the taste labels and dish labels of dishes, and calculate the recommendation value of each dish, which can well measure the user's preference for a certain dish.

[0035] As one of the implementation schemes, before making meal recommendations to the target user based on the target recommendation list, it also includes: determining the total number of meals consumed by the target user in the restaurant based on the target user's second historical consumption data, and judging whether the total number of consumed meals is greater than a preset evaluation threshold, and if so, making meal recommendations to the target user based on the target recommendation list, otherwise, making meal recommendations to the target user based on the target recommendation list and the meal recommendation set.

[0036] Specifically, the target recommendation list is determined based on the user's corresponding historical consumption data. The larger the data volume, the more accurate the corresponding result. However, when the data volume is small, the representativeness of the target recommendation list may decrease. Therefore, before recommending meals to the user, the total number of meals consumed in the restaurant can be analyzed first. After reaching a pre-set evaluation threshold, meal recommendations can be made to the user based on the target recommendation list. If the preset evaluation threshold is not reached, it means that the target recommendation list is not very targeted. In this case, a more appropriate recommendation plan can be generated in combination with the meal recommendation set.

[0037] As one of the implementation schemes, meal recommendations are made to the target user based on the target recommendation list and the meal recommendation set, including: filtering out a recommendation sub-list corresponding to each taste tag from the meal recommendation set based on taste tags corresponding to the meals included in all the orders of the target user, sorting the meals included in one or more filtered recommendation sub-lists according to the total recommendation list to obtain an auxiliary recommendation list, and collaboratively recommending the meals recorded in the target recommendation list and the auxiliary recommendation list to the target user.

[0038] Specifically, when the amount of sample data used for analysis is small, the flavors of the user's favorite dishes can be considered to determine the recommended sub-lists corresponding to the user's favorite flavors. Since the sample data is less representative, the priority between the flavors will not be considered for the time being. The dishes under each flavor are considered centrally, and the dishes under multiple flavors are sorted as a whole through the recommended total list to obtain an auxiliary recommendation list, which is used to assist the target recommendation list in making food recommendations to users when the amount of sample data is small.

[0039] Among them, the present application adopts a collaborative recommendation method of the target recommendation list and the auxiliary recommendation list, that is, while recommending a dish in the target recommendation list to the target user, a dish in the auxiliary recommendation list is recommended to the target user to improve the accuracy of the recommendation.

[0040] Example 2: Figure 2 FIG. 1 is a schematic diagram of a structure of a meal recommendation system based on big data analysis in an embodiment of the present application. Based on the embodiment 1, the embodiment 2 of the present application also provides a meal recommendation system based on big data analysis, including: An identity analysis module is used to obtain the identity information of the target user and determine whether the user is a new user based on the identity information; A data acquisition module, used to acquire the first historical consumption data of the restaurant and the second historical consumption data of the target user in the restaurant; A data analysis module, used to analyze the first historical consumption data of the restaurant to obtain a set of recommended dishes, and used to analyze the second historical consumption data of the target user in the restaurant to obtain a target recommendation list; The meal recommendation module is used to make meal recommendations to users based on the meal recommendation set and / or the target recommendation list.

[0041] As one implementation scheme, the data analysis module also includes: determining the total number of meals consumed by the target user in the restaurant based on the target user's second historical consumption data in the restaurant, and generating an auxiliary recommendation list by analyzing the target user's second historical consumption data in the restaurant and a set of meal recommendations.

[0042] It should be understood that those skilled in the art can make improvements or changes based on the above description, and all such improvements and changes should fall within the scope of protection of the appended claims of the present invention. Parts not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A meal recommendation method based on big data analysis, characterized in that: include: Obtain the identity information of the target user, determine whether the target user is a new user, and if so, recommend food to the target user based on the food recommendation set obtained by analyzing the first historical consumption data of the restaurant; otherwise, obtain the second historical consumption data of the target user in the restaurant, determine a target recommendation list based on the second historical consumption data, and recommend food to the target user based on the target recommendation list; The restaurant's meal recommendation set is determined in the following way: Get the restaurant's flavor label information and each dish's food label information; According to the first historical consumption data of the restaurant, the consumption times of each dish are counted respectively, and the dishes are sorted according to the consumption times to generate a total recommendation list; According to the taste label information and the food label information, a food collection with each taste label as the main body is established; For any set of dishes, all dishes in the set are sorted based on the number of times they are consumed, and a recommended sublist corresponding to the set of dishes is obtained; Traverse all meal sets, obtain the recommended sublist corresponding to each meal set, and construct a meal recommendation set based on the total recommendation list and multiple recommended sublists.

2. The method according to claim 1, characterized in that Recommend food to target users based on the food recommendation set, including: Obtaining user preference information, and filtering a first reference list and / or a second reference list from the meal recommendation set according to the preference information; The dishes recorded in the first reference list in the recommended total list are removed to obtain a third reference list, and the dishes recorded in the second reference list are reordered according to the recommended total list to obtain a fourth reference list; Make meal recommendations to the target user based on the third reference list and / or the fourth reference list.

3. The method according to claim 1, characterized in that Determine a target recommendation list based on the second historical consumption data, including: Extracting the food information included in each order in the second historical consumption data, determining the food label information of each order, and respectively counting the type of food labels included in each order and the first frequency proportion of each food label, as well as the taste labels corresponding to the food included in all orders and the second frequency proportion of each taste label; The priority of each food label is calculated using the following formula: , where is the priority of the food label, For food labels The first frequency percentage in orders, For frequency The normalization parameter of is the quantity of the order; After calculating the priority of each meal label, all meal labels included in the second historical consumption data are sorted according to the priority of each meal label to obtain a label list; Determine the weight parameter for each food label in the label list , for any food label , is the number of items in the food label that meet the condition , is the total number of food labels in the label list; According to the weight parameter of each food label in the label list , the recommended value of each dish in the restaurant is determined by the following formula: , where Indicates the recommended value. is the total number of food labels for the food, For the first Food labels, For the first The weight parameter of each food label, The second frequency ratio of the food’s taste label; Sort the dishes in the restaurant according to the recommendation value of each dish to obtain a target recommendation list.

4. The method according to claim 1, characterized in that Before making meal recommendations to the target user according to the target recommendation list, the method also includes: determining the total number of meals consumed by the target user in the restaurant according to the second historical consumption data of the target user, and judging whether the total number of consumed meals is greater than a preset evaluation threshold; if so, making meal recommendations to the target user according to the target recommendation list; otherwise, making meal recommendations to the target user according to the target recommendation list and the meal recommendation set.

5. The method according to claim 4, characterized in that Recommending meals to target users based on a target recommendation list and a meal recommendation set includes: filtering out a recommendation sub-list corresponding to each taste tag from the meal recommendation set based on taste tags corresponding to meals included in all orders of the target users, sorting meals included in one or more filtered recommendation sub-lists according to the total recommendation list to obtain an auxiliary recommendation list, and collaboratively recommending meals recorded in the target recommendation list and the auxiliary recommendation list to the target user.

6. The method according to claim 5, characterized in that Collaboratively recommending food to a target user includes: while recommending a food in a target recommendation list to the target user, recommending a food in an auxiliary recommendation list to the target user.

7. A meal recommendation system based on big data analysis, the system being applied to a meal recommendation method based on big data analysis as described in any one of claims 1 to 6, characterized in that: include: An identity analysis module is used to obtain the identity information of the target user and determine whether the user is a new user based on the identity information; A data acquisition module, used to acquire the first historical consumption data of the restaurant and the second historical consumption data of the target user in the restaurant; A data analysis module, used to analyze the first historical consumption data of the restaurant to obtain a set of recommended dishes, and used to analyze the second historical consumption data of the target user in the restaurant to obtain a target recommendation list; The meal recommendation module is used to make meal recommendations to users based on the meal recommendation set and / or the target recommendation list.

8. The system according to claim 7, characterized in that The data analysis module also includes: determining the total number of meals consumed by the target user in the restaurant based on the target user's second historical consumption data in the restaurant, and generating an auxiliary recommendation list by analyzing the target user's second historical consumption data in the restaurant and a set of meal recommendations.

Images