A catering information recommendation system based on big data

Abstract

The present invention relates to the field of auxiliary recommendation, and discloses a catering information recommendation system based on big data, including: a data collection module, used to collect historical dining information of target users; real-time collection of environmental information of target users and real-time needs of target users; a user habit analysis module, used to input historical dining information of target users into a trained recurrent neural network model, and output hot influencing factors affecting catering choices of target users; a merchant recommendation sequence generation module, which obtains the habit change rate of target users, compares the habit change rate of target users with a preset habit change interval, and divides target users into stable dining types and active dining types according to the comparison results; and selects different modes according to the dining type of target users to generate corresponding merchant recommendation sequences. The present invention not only improves the accuracy and user satisfaction of catering recommendations, but also realizes the perfect integration of user needs and personalized services.

Description

Technical Field

[0001] The present invention relates to the field of auxiliary recommendation, and in particular to a restaurant information recommendation system based on big data. Background Art

[0002] In the era of information overload, the catering industry has generated a large amount of data, including business information, user reviews, prices and tastes of dishes, etc.; users often find it difficult to filter out catering information that meets their needs from this massive amount of information. At the same time, with the improvement of living standards, users' demands for catering are becoming more and more diversified, including taste, price, environment, service and other aspects; traditional catering recommendation methods can no longer meet users' personalized needs; the catering industry is highly competitive, and businesses need to use effective recommendation systems to attract users and increase their popularity and sales.

[0003] With the continuous development of big data technology, the ability to collect, store, process and analyze data has been greatly improved; big data technology can realize the rapid processing and analysis of massive catering data, providing strong support for the recommendation system.

[0004] However, the existing restaurant information recommendation system still has some defects. For example, although it can generate a restaurant information recommendation list based on the user's historical information, in most cases it ignores the real-time needs of the target users, resulting in the inability to effectively match users with recommended merchants under different real-time needs. Not only does it fail to improve the efficiency of users selecting merchants, but it also increases the workload of users re-selecting merchants, which greatly affects the user experience. Summary of the invention

[0005] The purpose of the present invention is to provide a restaurant information recommendation system based on big data to solve the above technical problems.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A restaurant information recommendation system based on big data, comprising:

[0008] Data collection module, used to collect historical dining information of target users;

[0009] Collect the target user's environmental information and real-time needs in real time;

[0010] The user habit analysis module is used to input the target user's historical dining information into the trained recurrent neural network model and output the hot influencing factors that affect the target user's dining choices;

[0011] The merchant recommendation sequence generation module obtains the habit change rate of the target user, compares the habit change rate of the target user with the preset habit change interval, and divides the target user into stable dining type and active dining type according to the comparison result;

[0012] According to the target user's dining type, different modes are selected to generate corresponding merchant recommendation sequences.

[0013] A further solution is to obtain the habit change rate of the target user as follows:

[0014] According to formula (1): Calculate the habit change rate of target users ;

[0015] According to formula (2): Calculate the habit fluctuation index of target users ;

[0016] In the formula Indicates the current target user. Indicates the historical impact index of the target user's dining type. If the target user is an active dining type, If the target users are regular diners, ; represents the weight coefficient, Indicates the total number of preset cycles, Indicates Preset cycles, Indicates the total number of hotspot influencing factors, Indicates The first The frequency of occurrence of hot influencing factors, Indicates The frequency mean of the hotspot influencing factors, Indicates the average duration of each meal in each preset period. Indicates the average number of meals in each preset period. Indicates the duration of each preset cycle.

[0017] A further solution is to divide the target users into stable dining users and active dining users according to the comparison results:

[0018] If the target user's habit change rate , then the target users are active in dining;

[0019] If the target user's habit change rate , then the target users are the stable dining type; Indicates the preset habit change interval.

[0020] A further solution is to select different modes according to the target user's dining type to generate the merchant recommendation sequence:

[0021] When the target user is an active dining type, recommendation mode 1 is used to generate a merchant recommendation sequence; when the target user is a stable dining type, recommendation mode 2 is used to generate a merchant recommendation sequence.

[0022] In a further solution, the working process of the mode 1 is as follows:

[0023] The location of the target user is obtained based on GPS, and a circular area with a preset radius r with the location as the center is used as the target dining area;

[0024] Obtain environmental information of the target dining area and real-time needs of target users;

[0025] Process the environmental information to obtain the environmental coefficient , process the real-time demand of the target user to obtain the real-time demand coefficient ;

[0026] Substitute the environmental coefficient and the real-time demand coefficient into formula (3): Calculate the catering evaluation value of each merchant in the current area ; In the formula Indicates the distance between the location and each business;

[0027] Each business is evaluated according to the catering value The values ​​of are arranged in descending order, and the merchants in the first y positions are used as the merchant recommendation sequence for the target user.

[0028] In a further solution, the working process of the second mode is as follows:

[0029] The location of the target user is obtained based on GPS, and a circular area with a preset radius r with the location as the center is used as the target dining area;

[0030] Obtain information about each business in the target dining area, match the hotspot influencing factors of the target user with the information of each business, and score 1 point for each match;

[0031] Arrange each merchant in descending order according to the score, and put the merchants in the top y positions into the merchant recommendation queue;

[0032] Then according to the catering evaluation value of each merchant , after sorting the merchants in the merchant recommendation queue in descending order, they are used as the merchant recommendation sequence for the target user.

[0033] In a further embodiment, the environmental factor The expression is:

[0034] ;

[0035] In the formula, Represents the exponential coefficient, which is a constant. When the target user is an active dining user, ; When the target users are regular meal users, ; , Indicates the preset detection period. It indicates the temperature variation curve over time during the preset detection period. It represents the rainfall variation curve over time during the preset detection period. Indicates the number of preset detection periods. , Indicates the preset scale factor.

[0036] In a further embodiment, the real-time demand coefficient The expression is:

[0037] ;

[0038] In the formula , represents the weight coefficient, Indicates the hunger feeling value, Indicates the time until the next schedule. Indicates the maximum hunger value of the target user. Indicates the maximum time until the next schedule of the target user.

[0039] Beneficial effects of the present invention:

[0040] (1) The data collection module in the present invention can comprehensively and in real time capture the target user's historical dining records, current environment and immediate dining needs, laying a solid foundation for personalized recommendations; the user habit analysis module cleverly uses a well-trained recurrent neural network model to deeply analyze the user's historical dining data and accurately extract the hot factors that affect dining decisions, such as taste preferences, consumption levels, and dining time periods, thereby achieving a deep understanding of the user's dining preferences.

[0041] (2) The present invention further introduces the concept of habit change rate through the merchant recommendation sequence generation module. Through intelligent comparison with the preset interval, users are accurately divided into two categories: stable dining type and active dining type. For different user types, the system can flexibly switch recommendation modes to generate a merchant recommendation sequence that is highly consistent with user needs. For stable dining users, the system focuses on recommending their frequent or favorite restaurants; while for active dining users, it pays more attention to exploring novel and popular dining options. In summary, the present invention not only improves the accuracy and personalization of dining recommendations, but also realizes seamless connection between dining information and consumer needs through deep insight into user needs, bringing users a more convenient and pleasant dining experience, while also providing catering merchants with a more accurate and efficient marketing channel, achieving a win-win situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The present invention will be further described below in conjunction with the accompanying drawings.

[0043] Figure 1 It is a system structure block diagram of the present invention. DETAILED DESCRIPTION

[0044] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0045] See also Figure 1 As shown, the present invention is a restaurant information recommendation system based on big data, comprising:

[0046] Data collection module, used to collect historical dining information of target users;

[0047] Collect the target user's environmental information and real-time needs in real time;

[0048] The user habit analysis module is used to input the target user's historical dining information into the trained recurrent neural network model and output the hot influencing factors that affect the target user's dining choices;

[0049] The merchant recommendation sequence generation module obtains the habit change rate of the target user, compares the habit change rate of the target user with the preset habit change interval, and divides the target user into stable dining type and active dining type according to the comparison result;

[0050] The process of dividing the target users into stable dining types and active dining types according to the comparison results is as follows:

[0051] If the target user's habit change rate , then the target users are active in dining;

[0052] If the target user's habit change rate , then the target users are the stable dining type; Indicates the preset habit change interval;

[0053] Select different modes according to the target user's dining type to generate corresponding merchant recommendation sequences;

[0054] The process of selecting different modes to generate merchant recommendation sequences according to the target user's dining type is as follows:

[0055] When the target user is an active dining type, recommendation mode 1 is used to generate a merchant recommendation sequence; when the target user is a stable dining type, recommendation mode 2 is used to generate a merchant recommendation sequence.

[0056] The present invention, through the data collection module, can comprehensively and in real time capture the target user's historical dining records, current environment and immediate dining needs, laying a solid foundation for personalized recommendations; the user habit analysis module cleverly uses a well-trained recurrent neural network model to deeply analyze the user's historical dining data and accurately extract the hot factors that affect dining decisions, such as taste preferences, consumption levels, and dining time periods, thereby achieving a deep understanding of the user's dining preferences; the merchant recommendation sequence generation module further introduces the concept of habit change rate, and through intelligent comparison with the preset interval, accurately divides users into two types: stable dining type and active dining type. major categories; for different user types, the system can flexibly switch recommendation modes to generate a merchant recommendation sequence that is highly consistent with user needs; for stable dining users, the system focuses on recommending restaurants they often visit or like; and for active dining users, it pays more attention to exploring novel and popular dining options; in summary, the present invention not only improves the accuracy and personalization of dining recommendations, but also through deep insight into user needs, realizes the seamless connection between dining information and consumer needs, brings users a more convenient and pleasant dining experience, but also provides catering merchants with a more accurate and efficient marketing channel, achieving a win-win situation.

[0057] The process of obtaining the habit change rate of target users is:

[0058] According to formula (1): Calculate the habit change rate of target users ;

[0059] According to formula (2): Calculate the habit fluctuation index of target users ;

[0060] In the formula Indicates the current target user. Indicates the historical impact index of the target user's dining type. If the target user is an active dining type, If the target users are regular diners, ; Represents the weight coefficient, which is formulated based on historical data and empirical data. Indicates the total number of preset cycles, Indicates Preset cycles, Indicates the total number of hotspot influencing factors, Indicates The first The frequency of occurrence of hot influencing factors, Indicates The frequency mean of the hotspot influencing factors, Indicates the average duration of each meal in each preset period. Indicates the average number of meals in each preset period. Indicates the duration of each preset cycle.

[0061] The present invention provides a method for obtaining the habit change rate of target users, by obtaining the standard deviation formula between the frequency of occurrence and the mean of each hotspot influencing factor in multiple preset periods. Calculate the habit change rate of target users ; Then through the formula: Calculate the habit fluctuation index of target users , the degree of dispersion of each hotspot influencing factor is obtained by summing up. Obviously, if the degree of dispersion is larger, the rate of change of the hotspot influencing factor is larger and more unstable. On the contrary, if the degree of dispersion is smaller, the rate of change of the hotspot influencing factor is smaller and more stable. Therefore, the habit change rate of the target user is obtained by combining the degree of dispersion of each hotspot influencing factor, which provides a solid foundation for the subsequent classification of user dining types, and then generates different merchant recommendation sequences for users of different dining types to improve the effectiveness of the recommendation service; at the same time, combined with It can be seen that the current target user's dining frequency and dining duration. Obviously, the larger the value, the higher the target user's dining frequency and the longer the dining duration. Therefore, the habit change rate of the target user is calculated. The higher the accuracy.

[0062] The working process of Mode 1 is as follows:

[0063] The location of the target user is obtained based on GPS, and a circular area with a preset radius r with the location as the center is used as the target dining area;

[0064] Obtain environmental information of the target dining area and real-time needs of target users;

[0065] Process the environmental information to obtain the environmental coefficient , process the real-time demand of the target user to obtain the real-time demand coefficient ;

[0066] Substitute the environmental coefficient and the real-time demand coefficient into formula (3): Calculate the catering evaluation value of each merchant in the current area ; In the formula Indicates the distance between the location and each business;

[0067] The environmental factor The expression is:

[0068] ;

[0069] In the formula, Represents the exponential coefficient, which is a constant. When the target user is an active dining user, ; When the target users are regular meal users, ; , Indicates the preset detection period. It indicates the temperature variation curve over time during the preset detection period. It represents the rainfall variation curve over time during the preset detection period. Indicates the number of preset detection periods. , Represents the preset scaling factor, which is determined based on historical data analysis.

[0070] The real-time demand factor The expression is:

[0071] ;

[0072] In the formula , Represents the weight coefficient, which is formulated based on historical data and empirical data. Indicates the hunger feeling value, Indicates the time until the next schedule. Indicates the maximum hunger value of the target user. Indicates the maximum time until the next schedule of the target user;

[0073] The hunger feeling value is obtained through the hunger feeling scale, the scale range is usually from 0 to 10, or set to other ranges according to specific needs, such as 1 to 7; quantitative expression: individuals choose the corresponding number on the scale according to their feelings, and the larger the number, the higher the hunger level;

[0074] For example, if 5 is selected, it may indicate moderate hunger; if 10 is selected, it may indicate extreme hunger. The above-mentioned method of obtaining the hunger feeling value is prior art and will not be described in detail here.

[0075] Each business is evaluated according to the catering value The values ​​of are arranged in descending order, and the merchants in the first y positions are used as the merchant recommendation sequence for the target user.

[0076] The present invention is aimed at the target users who are active in dining. Since the target users are more random in choosing merchants, the environment and real-time needs of the target users are mainly considered. Calculate the environmental factor Obviously, if the temperature changes more during the detection period, it means that the temperature rises or drops sharply. At this time, the target user pays special attention to the distance of the dining place. Therefore, the closer the distance between the merchant and the target user, the more worthy of recommendation the merchant is, and the higher the probability of the target user adopting the recommendation is.

[0077] Also through the formula Calculate the real-time demand coefficient It is obvious that the more hungry the target user is, The larger the size, the more immediate the meal needs to be, and the shorter the time until the next schedule. The larger the value is, the shorter the user's dining time is and the more he needs to eat immediately. Therefore, the influence of dining distance increases exponentially. Finally, substitute it into the formula Calculate the catering evaluation value of each merchant in the current area The above formula can be used to accurately recommend effective merchants to active dining users, thereby improving user satisfaction.

[0078] The working process of the second mode is as follows:

[0079] The location of the target user is obtained based on GPS, and a circular area with a radius of r is preset with the location as the center as the target dining area; for example, the target user is currently located in a commercial district in the city center, with specific coordinates (longitude X, latitude Y), and a circular area with a radius of 500 meters with (longitude X, latitude Y) as the center is determined as the target dining area;

[0080] Obtain information about each merchant in the target dining area, match the target user's hotspot influencing factors with each merchant's information, and score 1 point for each match; obtain basic information about each merchant in the target dining area through a database or API interface, including name, address, cuisine, price range, and business hours; for example, there are 10 merchants in the target dining area, and the information of each merchant is recorded in detail; determine the target user's hotspot influencing factors, such as taste preferences, price sensitivity, and environmental requirements, based on their personal preferences, historical behaviors, and other data;

[0081] Assume that the target user prefers Sichuan cuisine, has a medium price sensitivity, and has certain requirements for the restaurant environment; among the 10 merchants, 5 offer Sichuan cuisine, 3 of which meet the user's requirements in terms of price, and 2 have a good environment; therefore, these 2 restaurants match the user in terms of taste preference, price sensitivity, and environment requirements, and each gets 3 points;

[0082] Arrange each merchant in descending order according to the score, and put the top y merchants, such as the top 5 merchants, into the merchant recommendation queue;

[0083] Then according to the catering evaluation value of each merchant , after sorting the merchants in the merchant recommendation queue in descending order, they are used as the merchant recommendation sequence for the target user.

[0084] The present invention, by setting the target dining area with the user's location as the center, can accurately lock the dining range that the user may be interested in, and then, by matching the user's hot spot influencing factors with the merchant information, can further screen out the merchants that match the user's preferences. In addition, through a double sorting mechanism - first sorting in descending order by matching scores, and then sorting in descending order by dining evaluation values, it can ensure that the recommended merchants not only meet the user's preferences but also have a high overall quality; this recommendation mode not only improves the accuracy of the recommendation, but also greatly enhances the user's dining experience, allowing users to quickly find high-quality restaurants that meet their needs.

[0085] It should be noted that the calculation formulas and various parameters involved in the calculation in the present invention have been dimensionally processed in advance, and the process of dimensionless processing is well known in the industry and will not be described here.

[0086] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A restaurant information recommendation system based on big data, characterized in that: include: Data collection module, used to collect historical dining information of target users; Collect the target user's environmental information and real-time needs in real time; The user habit analysis module is used to input the target user's historical dining information into the trained recurrent neural network model and output the hot influencing factors that affect the target user's dining choices; The merchant recommendation sequence generation module obtains the habit change rate of the target user, compares the habit change rate of the target user with the preset habit change interval, and divides the target user into stable dining type and active dining type according to the comparison result; Select different modes according to the target user's dining type to generate corresponding merchant recommendation sequences; The process of selecting different modes to generate merchant recommendation sequences according to the target user's dining type is as follows: When the target user is an active dining user, recommendation mode 1 is used to generate a merchant recommendation sequence; when the target user is a stable dining user, recommendation mode 2 is used to generate a merchant recommendation sequence; The working process of Mode 1 is as follows: The location of the target user is obtained based on GPS, and a circular area with a preset radius r with the location as the center is used as the target dining area; Obtain environmental information of the target dining area and real-time needs of target users; Process the environmental information to obtain the environmental coefficient , process the real-time demand of the target user to obtain the real-time demand coefficient ; Substitute the environmental coefficient and the real-time demand coefficient into formula (3): Calculate the catering evaluation value of each merchant in the current area ; In the formula Indicates the distance between the location and each business; Each business is evaluated according to the catering value The values ​​of are sorted in descending order, and the merchants in the top y positions are used as the merchant recommendation sequence for the target user; The working process of the second mode is as follows: The location of the target user is obtained based on GPS, and a circular area with a preset radius r with the location as the center is used as the target dining area; Obtain information about each business in the target dining area, match the hotspot influencing factors of the target user with the information of each business, and score 1 point for each match; Arrange each merchant in descending order according to the score, and put the merchants in the top y positions into the merchant recommendation queue; Then according to the catering evaluation value of each merchant , after sorting the merchants in the merchant recommendation queue in descending order, they are used as the merchant recommendation sequence for the target user.

2. The restaurant information recommendation system based on big data according to claim 1 is characterized in that: The process of obtaining the habit change rate of target users is: According to formula (1): Calculate the habit change rate of target users ; According to formula (2): Calculate the habit fluctuation index of target users ; In the formula Indicates the current target user. Indicates the historical impact index of the target user's dining type. If the target user is an active dining type, If the target users are regular diners, ; represents the weight coefficient, Indicates the total number of preset cycles, Indicates Preset cycles, Indicates the total number of hotspot influencing factors, Indicates The first The frequency of occurrence of hot influencing factors, Indicates The frequency mean of the hotspot influencing factors, Indicates the average duration of each meal in each preset period. Indicates the average number of meals in each preset period. Indicates the duration of each preset cycle.

3. The restaurant information recommendation system based on big data according to claim 2 is characterized in that: The process of dividing the target users into stable dining types and active dining types according to the comparison results is as follows: If the target user's habit change rate , then the target users are active in dining; If the target user's habit change rate , then the target users are the stable dining type; Indicates the preset habit change interval.

4. The restaurant information recommendation system based on big data according to claim 1, characterized in that: The environmental factor The expression is: ; In the formula, Represents the exponential coefficient, which is a constant. When the target user is an active dining user, ; When the target users are regular meal users, ; , Indicates the preset detection period. It indicates the temperature variation curve over time during the preset detection period. It indicates the rainfall variation curve over time during the preset detection period. Indicates the number of preset detection periods. , Indicates the preset scale factor.

5. The restaurant information recommendation system based on big data according to claim 1 is characterized in that: The real-time demand factor The expression is: ； In the formula , represents the weight coefficient, Indicates the hunger feeling value, Indicates the time until the next schedule. Indicates the maximum hunger value of the target user. Indicates the maximum time until the next schedule of the target user.

Images