Dish recommendation system, dish recommendation method, and program

Abstract

The dish recommendation system according to the present invention comprises: an acquisition unit for pre-acquiring a plurality of event data in which identification information, dish category information, and eating time information of dishes eaten by a user are classified by dish category information, the dish category information including taste feeling information in which the dishes are classified by taste feelings; and an extraction unit that calculates a plurality of characteristic indexes by time-series correlation analysis on the basis of the plurality of event data, extracts a specific characteristic index obtained by analyzing the dietary habit of the user from the plurality of characteristic indexes on the basis of specific event data corresponding to specific dish category information input by the user, and extracts the extracted specific characteristic index from the plurality of characteristic indexes. The plurality of characteristic indexes are respectively calculated for each piece of dish category information and comprise at least one characteristic index of support degree, confidence degree and improvement degree related to the dietary habits of the user; and a recommendation unit that outputs dish candidates proposed to the user on the basis of the specific characteristic index.

Description

Technical Field

[0001] The present invention relates to a dish recommendation system, a dish recommendation method and a program. Background Art

[0002] In the past, a dish recommendation system for proposing dishes based on a user's eating history has been proposed. For example, Patent Document 1 discloses a dish recommendation device for proposing a dish to be eaten next using an eating pattern, preference, and food intake, wherein the eating pattern is an eating pattern based on past eating history.

[0003] Prior art literature

[0004] Patent Literature

[0005] Patent Document 1: Japanese Patent Application Publication No. 2022-124701 Summary of the invention

[0006] Problem that the invention aims to solve

[0007] However, since the device of Patent Document 1 automatically proposes dishes without confirming the user's request, there is a possibility that the proposed dishes are different from the dishes that the user wants to eat.

[0008] The purpose of the present disclosure is to provide a dish recommendation system, a dish recommendation method and a program for proposing suitable dishes to users.

[0009] Solutions to the problem

[0010] The dish recommendation system disclosed in the present invention includes: an acquisition unit, which pre-acquires a plurality of event data that classifies identification information, dish category information and eating time information of dishes that the user has eaten according to the dish category information, wherein the dish category information includes taste sensation information that classifies the dishes according to the sensation of taste; an extraction unit, which calculates a plurality of characteristic indicators through time series association analysis based on the plurality of event data, and extracts a specific characteristic indicator obtained by analyzing the user's eating habits from the plurality of characteristic indicators based on specific event data corresponding to the specific dish category information input by the user, the plurality of characteristic indicators being characteristic indicators calculated separately for each dish category information and including at least one of support, confidence and lift related to the user's eating habits; and a recommendation unit, which outputs dish candidates proposed to the user based on the specific characteristic indicators.

[0011] The dish recommendation method disclosed in the present invention includes the following steps: pre-acquiring a plurality of event data that classifies identification information of dishes that the user has eaten, dish category information, and eating time information according to the dish category information, wherein the dish category information includes taste sensation information that classifies dishes according to taste sensation; based on the plurality of event data, calculating a plurality of characteristic indicators through time series association analysis, and based on specific event data corresponding to the specific dish category information input by the user, extracting a specific characteristic indicator obtained by analyzing the user's eating habits from the plurality of characteristic indicators, wherein the plurality of characteristic indicators are characteristic indicators calculated separately for each dish category information and include at least one of support, confidence, and lift related to the user's eating habits; and outputting dish candidates proposed to the user based on the specific characteristic indicators.

[0012] The program disclosed in the present invention enables a computer to perform the following processing: pre-acquire a plurality of event data that classifies identification information, dish category information, and eating time information of dishes that a user has eaten according to the dish category information, wherein the dish category information includes taste sensation information that classifies dishes according to the sensation of taste; based on the plurality of event data, calculate a plurality of characteristic indicators through time series association analysis, and based on specific event data corresponding to specific dish category information input by the user, extract a specific characteristic indicator obtained by analyzing the user's eating habits from the plurality of characteristic indicators, wherein the plurality of characteristic indicators are characteristic indicators calculated separately for each dish category information and include at least one of support, confidence, and lift related to the user's eating habits; and based on the specific characteristic indicators, output dish candidates proposed to the user.

[0013] Effects of the Invention

[0014] According to the present disclosure, suitable dishes can be suggested to users. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is a diagram showing an outline of one embodiment of the present disclosure.

[0016] Figure 2 1 is a block diagram showing the functional structure of a dish recommendation system according to one embodiment of the present disclosure.

[0017] Figure 3 1 is a block diagram showing the hardware structure of a dish recommendation system according to one embodiment of the present disclosure.

[0018] Figure 4 It is a flowchart showing the operation of one embodiment of the present disclosure.

[0019] Figure 5 This is a diagram showing dish category information determined based on ingredient information.

[0020] Figure 6 is a graph representing event data.

[0021] Figure 7 This is a diagram showing analysis data obtained by performing a time-series correlation analysis on user event data.

[0022] Figure 8 This is a diagram showing analysis data obtained by normalizing characteristic indices.

[0023] Fig. 9 This is a diagram showing a state of performing time series correlation analysis on event data of other users.

[0024] Fig.10 This is a graph showing comparative data obtained by comparing the characteristic indexes of the user with those of other users.

[0025] Fig.11 This is a diagram showing suggestion data including menu candidates output from the recommendation unit.

[0026] Fig.12 This is a diagram showing the eating habit data output from the recommendation unit. DETAILED DESCRIPTION

[0027] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0028] The summary of the present disclosure is described. Figure 1As shown, if the eating data (such as images of dishes, etc.) that the user has eaten is input, the user's terminal 1 sends the eating data to the server 2 in sequence. Here, it is assumed that the server 2 has the dish recommendation system disclosed in the present invention. If the dish recommendation system of the server 2 receives the eating data from the terminal 1, it obtains the identification information (such as name, etc.) of the dishes that the user has eaten, the dish category information (such as light, etc.) and the eating time information (such as date, etc.) based on the eating data, and pre-produces a plurality of event data that classify these information according to the dish category information. In addition, the dish recommendation system analyzes the characteristics of the user's eating habits based on the event data. At this time, if the user's eating habits are analyzed based on the event data containing only the identification information of the dishes and the eating time information, the eating habits of dishes containing various categories are analyzed, and it is difficult to analyze the eating habits based on the dish category (for example, the eating habits of "light" dishes, etc.). Here, when specific dish category information (e.g., light, etc.) that the user wants to eat is input into the terminal 1, the terminal 1 sends the specific dish category information to the server 2. If the dish recommendation system of the server 2 receives the specific dish category information from the terminal 1, it extracts and analyzes the characteristics of the user's eating habits based on the specific event data corresponding to the specific dish category information. Then, the dish recommendation system outputs a dish candidate proposed to the user based on the extracted characteristic indicator, and sends the dish candidate to the terminal 1. As a result, the terminal 1 displays the dish candidate corresponding to the specific dish category information on the display unit.

[0029] Next, the structure of the dish recommendation system disclosed in the present invention is described in detail. Figure 2 As shown, the menu recommendation system 3 includes an acquisition unit 4, a storage unit 5, an extraction unit 6, and a recommendation unit 7. The acquisition unit 4 is connected to the storage unit 5, and the storage unit 5 is connected to the recommendation unit 7 via the extraction unit 6.

[0030] The acquisition unit 4 acquires a plurality of event data that classifies the identification information, the dish category information, and the eating time information of the dishes that the user has eaten according to the dish category information. Here, it can be that the dishes include single dishes (such as hamburger steaks, etc.), and menus representing multiple dishes (such as rice and hamburger steaks, etc.). In addition, as the identification information of the dishes, for example, the name of the dish, the identification number, etc. can be cited. In addition, the dish category information is information that classifies the dishes by category, for example, it can include taste information that classifies the dishes by the taste feeling. Here, as the taste feeling information, for example, the mouthfeel information representing the physical taste feeling, the taste information representing the chemical taste feeling, etc. can be cited. Specifically, the taste feeling information can include information representing onomatopoeia, that is, information representing onomatopoeia or onomatopoeia, which is an onomatopoeia that expresses the feeling of taste. As onomatopoeia, for example, light, strong, elastic, chewy, soft, crisp, sticky, etc. can be cited. Here, the information representing the onomatopoeia can be composed of at least a part of the onomatopoeia. In addition, the eating time information is information related to the time of eating, and the eating time information can include the time, date, time section (e.g., breakfast, lunch, dinner) of eating.

[0031] For example, the acquisition unit 4 can receive the eating data input by the user from the terminal 1. Here, the eating data is data related to the dishes that the user has eaten, such as images (photos, etc.) of the dishes, menus, questionnaires about the dishes (for example, preference information of the dishes, avoidance of ingredients information, attribute information of the user, etc.), and dish category information. The preference information of the dishes is information related to the dishes preferred by the user, such as the names of the dishes or the names of the ingredients preferred by the user. In addition, the avoidance of ingredients information is information related to the ingredients that the user avoids using in the dishes, such as ingredients that the user does not like, ingredients that have allergic reactions, etc. In addition, the attribute information of the user can include gender, whether the user is losing weight, whether there is a chronic disease (such as high blood pressure, etc.). If the acquisition unit 4 receives the eating data, it obtains the identification information of the dishes that the user has eaten, the dish category information and the eating time information based on the eating data, and creates a plurality of event data that classify these information according to the dish category information.

[0032] The storage unit 5 stores a plurality of event data acquired by the acquisition unit 4. That is, the storage unit 5 stores a plurality of event data that establishes a correspondence between the identification information of the dishes that the user has eaten and the eating time information and is classified according to the dish category information. In addition, the storage unit 5 outputs specific event data corresponding to the specific dish category information input by the user among the plurality of event data to the extraction unit 6. For example, when the acquisition unit 4 receives the specific dish category information input by the user to the terminal 1 from the terminal 1, the storage unit 5 outputs the specific event data corresponding to the specific dish category information to the extraction unit 6. For example, if the specific taste sensation information "light" is input by the user as the specific dish category information, the storage unit 5 outputs the event data corresponding to the "light" to the extraction unit 6.

[0033] Based on the multiple event data, the extraction unit 6 calculates multiple characteristic indicators through time series association analysis. The multiple characteristic indicators are characteristic indicators calculated for each of the dish category information and include at least one of the support, confidence and lift related to the user's eating habits. In addition, the extraction unit 6 extracts and analyzes the characteristics of the user's eating habits from the multiple characteristic indicators based on the specific event data input from the storage unit 5. Here, the characteristics of eating habits represent the tendency of users to choose dishes in multiple meals (for example, rules or conditions, etc.). For example, it can represent the tendency of users to choose dishes when eating "light" dishes three times. In addition, the characteristic indicator is an indicator that represents the characteristics of eating habits, for example, it can be calculated by comparing the event data of multiple meals.

[0034] The recommendation unit 7 outputs the dish candidates proposed to the user based on the characteristic index obtained by the extraction unit 6. At this time, the recommendation unit 7 can output the dish candidates based on the event data that satisfies at least one of the three conditions of "support is above the specified threshold", "confidence is above the specified threshold", and "lift is above the specified threshold". In addition, the recommendation unit 7 can also analyze the characteristics of the eating habits of other users for each dish category information based on a plurality of event data in which the identification information, dish category information, and eating time information of the dishes that other users have eaten are classified according to the dish category information. Then, the recommendation unit 7 can search for the dish candidates proposed to the user based on the characteristic index of the user calculated by the extraction unit 6 and the characteristic index of other users corresponding to the specific dish category information.

[0035] Figure 3 It shows the hardware structure of the dish recommendation system 3.

[0036] The menu recommendation system 3 includes a storage device 11 , a processor 12 , a user interface (UI) device 13 , and a communication device 14 , which are interconnected via a bus B.

[0037] In addition, the programs or instructions for implementing the various functions and processes described later in the dish recommendation system 3 can be downloaded from an external device via the network, etc., or provided by a removable storage medium such as a CD-ROM (Compact Disk-Read Only Memory) or a flash memory.

[0038] The storage device 11 is implemented by one or more non-transitory storage media such as a random access memory, a flash memory, a hard disk drive, etc., and stores installed programs or instructions and files and data used in executing the programs or instructions.

[0039] The processor 12 may be implemented by one or more CPUs (Central Processing Units) that may be composed of one or more processor cores, GPUs (Graphics Processing Units), processing circuits, etc. The processor 12 executes various functions and processes of the dish recommendation system 3 described later according to the program, instructions, and data such as parameters required to execute the program or instructions stored in the storage device 11.

[0040] The UI device 13 may be composed of input devices such as a keyboard, a mouse, a camera, a microphone, output devices such as a display, a speaker, a headset, a printer, and input and output devices such as a smart phone, a tablet computer, and a touch panel, to realize an interface between the manager and the dish recommendation system 3. For example, the manager may use a keyboard, a mouse, etc. to operate a GUI (Graphical User Interface) displayed on a display or a touch panel to operate the dish recommendation system 3.

[0041] In addition, the above hardware structure is only an example, and the dish recommendation system 3 of the present disclosure can also be implemented by any other appropriate hardware structure.

[0042] Next, refer to Figure 4 The flowchart shown explains the operation of this embodiment.

[0043] First, in step S1, Figure 2The acquisition unit 4 of the dish recommendation system 3 shown acquires event data including identification information, dish category information and eating time information of the dish that the user has eaten. For example, the acquisition unit 4 can acquire event data based on the user's eating data. For example, if the user inputs eating data such as a photographed image of the dish that has been eaten into the terminal 1, the terminal 1 sends the eating data to the server 2 in sequence. Then, in the server 2, when the acquisition unit 4 of the dish recommendation system 3 receives the user's eating data from the terminal 1, it acquires the identification information, dish category information and eating time information of the dish that the user has eaten based on the eating data. At this time, the acquisition unit 4 can also further acquire the food information of the dish that the user has eaten based on the eating data. In addition, the food information is the information of the food used in the dish, for example, it can include information such as ingredients (such as white radish, carrots, etc.), seasonings (such as soy sauce, sauce, etc.), and the amount of food.

[0044] For example, the acquisition unit 4 can identify the name (identification information) and ingredient information of the dish that the user has eaten based on the photographed image of the dish received as the eating data, and estimate the taste sensation information (dish category information) that classifies the dish by taste sensation based on the ingredient information. Figure 5 When the same dishes such as chicken stew 1 to chicken stew 5 are identified as shown, the acquisition unit 4 may also identify the differences in the food information of the food materials used in the chicken stews 1 to 5, and estimate the taste sensory information based on the food information. For example, when the acquisition unit 4 recognizes that the food information of chicken stew 1 is yam based on the captured image of chicken stew 1, the taste sensory information is estimated to be "soft" based on the taste of the yam. In addition, although the food information of chicken stew 4 and chicken stew 5 both show "white radish", they can be classified as "light" and "sticky" respectively based on other ingredients. That is, the taste sensory information can be comprehensively estimated based on the food information of multiple ingredients used in the dishes. In addition, the acquisition unit 4 can, for example, acquire the eating time information based on the shooting time attached to the eating data received from the terminal 1.

[0045] In addition, in the above-mentioned embodiment, the acquisition unit 4 acquires the identification information, dish category information, eating time information and ingredient information of the dish that the user has eaten based on the user's eating data, but is not limited to this, as long as this information can be acquired. For example, the acquisition unit 4 can also directly acquire the identification information, dish category information, eating time information and ingredient information of the dish that the user has eaten from the terminal 1. That is, it can also be that the terminal 1 receives the input of the identification information, dish category information, eating time information and ingredient information of the dish that the user has eaten from the user, and sends the identification information, dish category information, eating time information and ingredient information of the dish input by the user to the server 2.

[0046] Next, in step S2, the acquisition unit 4 generates a plurality of event data that classifies the acquired dish identification information, dish category information, and eating time information according to the dish category information, and stores the event data in the storage unit 5. At this time, the acquisition unit 4 may store a plurality of event data that establishes a correspondence between at least one of the dish and ingredient information that the user has eaten and the dish category information in the storage unit 5. For example, Figure 6 As shown, the acquisition unit 4 can create a plurality of event data 15a, 15b... that classify the name (identification information), taste information (dish category information) and eating time information of the dish according to the taste information "light", "strong taste"... Specifically, when the acquisition unit 4 acquires the data of April 1, it records the data in the event data 15a based on the taste information "light", and when the acquisition unit 4 acquires the data of April 2, it records the data in the event data 15b based on the taste information "strong taste". In addition, the acquisition unit 4 can create a plurality of event data 15a, 15b... that establish a correspondence between the name, taste information and food information of the dish that the user has eaten, according to the date shown in the eating time information, and store the plurality of event data 15a, 15b... in the storage unit 5. In this way, in each of the plurality of event data 15a, 15b..., the data of the same taste information is recorded in the order of eating time. Every time the eating data of the user is received, the acquisition unit 4 updates the event data 15a, 15b, ..., and constructs a plurality of event data 15a, 15b, ... related to the dishes eaten by the user in the past.

[0047] On the other hand, the terminal 1 may also accept a proposal commission from the user, which is a commission for proposing a dish that the user is going to start cooking from now on. At this time, the terminal 1 may accept input of specific dish category information. For example, the terminal 1 may accept specific taste sensation information that classifies dishes by taste sensation. At this time, the terminal 1 may accept information representing onomatopoeia as taste sensation information, which onomatopoeia is an onomatopoeia that expresses the sensation of taste. When specific dish category information is input by the user, the terminal 1 sends the specific dish category information to the server 2. Here, it is assumed that the user inputs the onomatopoeia "light" as specific taste sensation information into the terminal 1.

[0048] As described above, the dish category information includes taste sensation information for classifying dishes by the sensation of taste. For example, if dishes are classified by food culture such as "French dishes", it will be difficult to appropriately represent the dishes desired by the user because there are various categories of dishes in this category (for example, "light" dishes and "strong" dishes). Here, the user specifies the category of the dish using the taste sensation information, thereby being able to appropriately specify the range of the desired dishes. In addition, the taste sensation information includes information representing onomatopoeia, which is an onomatopoeia that expresses the sensation of taste. Therefore, the user can easily specify the taste sensation information.

[0049] In the dish recommendation system 3 of the server 2, if the specific taste sensation information "light" input by the user is received from the terminal 1, then in step S3, the extraction unit 6 selects the specific event data 15a corresponding to the specific taste sensation information "light" from the multiple event data 15a, 15b... stored in the storage unit 5. At this time, it can be that the extraction unit 6 selects the data of the specified time period (for example, 14 days) in the specific event data 15a. Then, in step S4, the extraction unit 6 can calculate the characteristic index representing the user's eating habits by performing a time series association analysis on the extracted specific event data 15a. As the characteristic index, for example, support, confidence, and lift for the combination of events can be cited. For example, the extraction unit 6 can calculate the characteristic index including at least one of support, confidence, and lift. In addition, it can also be that the extraction unit 6 automatically calculates the characteristic index for each dish category information based on the multiple event data 15a, 15b... stored in the storage unit 5 without receiving the specific taste sensation information from the user.

[0050] Specifically, it can be, Figure 6 As shown, the extraction unit 6 identifies the name of the dish, taste sensation information, and ingredient information in each column (for example, each day) as event data 15a1, event data 15a2, ... in the specific event data 15a of 14 days (prescribed time period), and then calculates the characteristic index for the combination of two event data contained in a preset analysis time period (for example, 3 days). For example, the extraction unit 6 uses April 10 as the base date, and sets April 10, April 8 (the last time), and April 5 (the last time) as the analysis time period, and calculates the characteristic index for the combination of two event data contained in the analysis time period. For example, the event before the base date is regarded as event A, and the event on the base date is regarded as event B, as shown in FIG. Figure 7As shown, the extraction unit 6 calculates the characteristic index 17 for each combination of the event data 15z1 belonging to the A event (the event on April 8 or April 5) and the event data 15z2 belonging to the B event (the event on April 10). Next, the extraction unit 6 shifts the date by one and sets April 8 as the reference date, and sets the three days of April 8, April 5 (the previous time), and April 3 (the previous time) as the next analysis time period. Similarly, the extraction unit 6 calculates the characteristic index 17 for each combination of the two event data 15z1 and 15z2 included in the analysis time period. In this way, the extraction unit 6 sequentially shifts the analysis time period in 14 days (predetermined time period) and calculates the characteristic index 17 for the combination of the two event data 15z1 and 15z2 included in each analysis time period. Then, the extraction unit 6 generates the analysis data 16 for 14 days (predetermined time period), and the analysis data 16 includes the event data 15z1 of the A event, the event data 15z2 of the B event, and the characteristic index 17.

[0051] Here, the support degree, for example, represents the number of combinations of event data 15z1 of event A and event data 15z2 of event B, and can be calculated using the following formula (1). In addition, the confidence degree, for example, represents the probability of consuming the dish of event A last time or the previous previous time before consuming the dish of event B, and can be calculated using the following formula (2). In addition, the lift degree, for example, represents the increase in the probability of consuming the dish of event A last time or the previous previous time before consuming the dish of event B, and can be calculated based on the following formula (3). In addition, the total number of data can be calculated, for example, as follows: the number of dish identification information × the number of dish category information × the number of ingredient information × the number of analysis time periods (for example, 3 days).

[0052] Support = the number of data representing the same combination of event data 15z1 of event A and event data 15z2 of event B ÷ the total number of data...(1)

[0053] Confidence = the number of data indicating that event data 15z1 of event A and event data 15z2 of event B are in the same combination ÷ the number of data including event data 15z2 of event B...(2)

[0054] Lift = the number of data representing the same combination of event data 15z1 of event A and event data 15z2 of event B ÷ the number of data including event data 15z2 of event B ÷ the number of data including event data 15z1 of event A ÷ the total number of data...(3)

[0055] At this time, the extraction unit 6 may also search the analysis data 16a based on the temporal sequence of the event data 15a1, 15a2, ... in the specific event data 15a, and filter out the data of the sequence that is not included in the specific event data 15a. For example, when event data 15z2 (equivalent to event B) is recorded after event data 15z1 (equivalent to event A) in the specific event data 15a, the extraction unit 6 may filter out the data indicating that event A is event data 15z2 and event B is event data 15z1 (data in the opposite order to the specific event data 15a) in the analysis data 16a. Thus, the specific event data 15a is subjected to association analysis in accordance with the time sequence.

[0056] In this way, the extraction unit 6 extracts and analyzes the characteristics of the user's eating habits based on the specific event data 15a corresponding to the specific dish category information "light" input by the user. At this time, since the plurality of event data 15a, 15b, ... classify the information of the dishes that the user has eaten according to the dish category information, the extraction unit 6 only needs to analyze the specific event data 15a corresponding to the specific dish category information to easily extract the characteristics of the user's eating habits.

[0057] Furthermore, the extracting unit 6 analyzes the characteristics of the eating habits of the user P by performing time series correlation analysis on the specific event data 15 a and outputs characteristic indexes 17 such as support, confidence, and lift. Thus, the extracting unit 6 can accurately indicate the eating habits of the user P using the characteristic indexes 17 .

[0058] In addition, the plurality of event data 15a, 15b, ... include the food information of the dishes eaten by the user P. Thus, the extracting unit 6 can analyze the characteristics of the eating habits of the user P more accurately based on the differences in the food information.

[0059] The plurality of event data 15a, 15b, ... are data formed by associating at least one of the information on dishes and ingredients eaten by the user P with the dish category information. Thus, the extraction unit 6 can easily analyze the characteristics of the user P's eating habits.

[0060] In addition, the dish classification information includes taste information for classifying dishes by taste. Since the extraction unit 6 performs analysis based on the taste information that has a greater impact on the user's eating habits, the characteristics of the eating habits of the user P can be accurately analyzed. In addition, the taste information includes information indicating onomatopoeia and mimetic words that express the taste. Since the extraction unit 6 performs analysis based on onomatopoeia and mimetic words that accurately express the taste information, the characteristics of the eating habits of the user P can be more accurately analyzed.

[0061] In addition, the extraction unit 6 may also use at least one of the preference information and the avoidance food information of the user P to extract the characteristic index 17. For example, the extraction unit 6 may also retrieve the analysis data 16 based on the preference information of the user P, and preferentially extract the event data including the characteristic index 17, which is the characteristic index 17 in which the dish or food corresponding to the preference information is registered. In addition, the extraction unit 6 may also retrieve the analysis data 16 based on the avoidance food information of the user P, and extract the event data including the characteristic index 17, which is the characteristic index 17 in which no avoidance food information is registered. That is, the extraction unit 6 may delete the event data in which the avoidance food information is registered from the analysis data 16. In addition, the preference information and the avoidance food information may be calculated based on the eating data of the user P, or may be set by the user P. Thus, the extraction unit 6 can appropriately extract the characteristic index 17 according to the preference information or avoidance food information of the user P.

[0062] In addition, the extraction unit 6 may further use the attribute information of the user P to extract the characteristic index 17. For example, when the attribute information of the user P is set to "diet", the extraction unit 6 may extract the event data of dishes with calories below a predetermined value. In addition, the attribute information may be set by the user P.

[0063] In addition, the extraction unit 6 may also use the cooking recipe browsing history to extract the characteristic index 17. For example, when the user P searches for the cooking recipe of a dish on the Internet, the terminal 1 may send the cooking recipe browsing information to the server 2. The extraction unit 6 may calculate the preference information of the user P based on the cooking recipe browsing history sent from the terminal 1. Furthermore, the extraction unit 6 may also preferentially extract event data including the characteristic index 17, which is a characteristic index 17 of a dish or ingredient corresponding to the calculated preference information of the user P.

[0064] In addition, the extraction unit 6 may also normalize the characteristic index. Figure 8 As shown, the extraction unit 6 can produce analysis data 16a in which the support, confidence, and lift of the characteristic index 17 are normalized and shown as variable values. In addition, the variable name is the name of the combination of the event data 15z1 of the event A and the event data 15z2 of the event B. Here, for the same combination of the event data 15z1 of the event A and the event data 15z2 of the event B, the support, confidence, and lift are separately represented in the form of variable names X, Y, and Z, that is, they are represented separately as support (X1, X2...), confidence (Y1,...), and lift (Z1,...).

[0065] Next, the extraction unit 6 outputs the extracted analysis data 16 including the characteristic index 17 to the recommendation unit 7. Here, it is assumed that the extraction unit 6 outputs the analysis data 16a to the recommendation unit 7. After the analysis data 16a is input, in step S5, the recommendation unit 7 compares the analysis data 16a of the user with the analysis data obtained by performing a time series correlation analysis on the characteristics of the eating habits of other users. Here, the recommendation unit 7 can also analyze the characteristics of the eating habits of other users according to the dish category information based on a plurality of event data in which the identification information, dish category information and eating time information of the dishes eaten by other users are classified according to the dish category information, similarly to the analysis of the user, and pre-store the obtained plurality of analysis data in the storage unit.

[0066] For example, Fig. 9 As shown, the recommendation unit 7 pre-acquires event data 18a, 18b, ... of multiple other users P1, P2, .... At this time, the recommendation unit 7 can analyze and produce multiple event data that classify the identification information, dish category information and eating time information of the dishes according to the dish category information for the other users P1, P2, .... For example, if the situation of the other user P1 is explained, the recommendation unit 7 can produce multiple event data 18a1, 18a2, ... that classify the identification information, taste sensation information and eating time information of the dishes that the other user P1 has eaten according to the taste sensation information "light", "strong taste" .... Then, based on each event data 18a1, 18a2 ..., the recommendation unit 7 analyzes the characteristics of the eating habits of the other users for each of the taste sensation information "light", "strong taste" .... At this time, the recommendation unit 7 can perform time series correlation analysis on the event data 18a1, 18a2 ... respectively to calculate the characteristic index 19 of the eating habits of the other users. Thus, the recommendation unit 7 generates analysis data 21a1, 21a2, ... each including event data 20z1 of event A, event data 20z2 of event B, and characteristic index 19. Furthermore, the analysis data 21a1, 21a2, ... are generated in correspondence with the taste sensation information "light", "strong", ... respectively. The recommendation unit 7 similarly calculates the characteristic index 19 representing the eating habits for other users P2, P3, ..., and generates analysis data 21b, 21c, ... including event data 20z1 of event A, event data 20z2 of event B, and characteristic index 19. The recommendation unit 7 stores the analysis data 21a, 21b, ... of other users P1, P2, ... in the storage unit in advance.

[0067] In addition, when the variable names of other users P1, P2, ... are inconsistent (the variable name of a specific user is missing), the recommendation unit 7 may also interpolate the missing variable name of the specific user based on the average value of the variable values ​​of other users P1, P2, ... whose variable names exist. In addition, the recommendation unit 7 may obtain the preference information, avoidance food information, attribute information, or cooking recipe browsing history of the user P from the extraction unit 6, and use this information to extract the characteristic index 19 in the same way as the extraction unit 6. For example, between step S5 and step S6, the recommendation unit 7 may filter the analysis data 21a, 21b, ... based on the preference information, avoidance food information, attribute information, or cooking recipe browsing history of the user P.

[0068] When the analysis data 16a of the user indicating the taste sensation information "light" is input from the extraction unit 6, the recommendation unit 7 selects the analysis data 21a1, 21b1, ... corresponding to the taste sensation information "light" from the analysis data 21a, 21b, ... of other users P1, P2, ... stored in the storage unit. Then, the recommendation unit 7 compares the analysis data 21a1, 21b1, ... corresponding to the taste sensation information "light" with the analysis data 16a of the user.

[0069] At this time, the recommendation unit 7 can search for dish candidates to be proposed to the user based on the characteristic index 17 indicating the eating habits corresponding to the specific taste sensation information "light" input by the user, which includes the analysis data 21a1, 21b1, ... of the eating habits of other users corresponding to the taste sensation information "light". Fig.10As shown, collaborative filtering is used to compare the variable value 23 (characteristic index) of user P with the variable values ​​24 (characteristic index) of other users P1, P2, ... in the comparison data 22, thereby searching for dish candidates proposed to user P. At this time, the recommendation unit 7 can calculate the similarity of the variable values ​​24 of other users P1, P2, ... relative to the variable value 23 of user P for all variables X1, X2, ..., Y1, ..., Z1, .... As an example, the recommendation unit 7 can also calculate the cosine similarity based on the vector of the variable value 23 of user P and the vector of the variable value 24 of other users P1, P2, .... In addition, the recommendation unit 7 selects other users whose similarity is greater than a specified value (for example, other users P1 and P2 whose similarity is greater than 0.8) as users with high similarity in eating habits with user P. Next, the recommendation unit 7 counts the number of variable values ​​24a that present values ​​within a specified value (e.g., the same value) relative to the variable value 23 of the user P, among the variable values ​​24 of other users P1 and P2 with similar eating habits, for each variable X1, X2, ..., The recommendation unit 7 then selects the variables X2, Y1, and Z1 having a large number of variable values ​​24a, and in step S6, outputs the dish candidates based on the selected variables X2, Y1, and Z1. At this time, the recommendation unit 7 may also preferentially select, from the selected variables X2, Y1, and Z1, a variable having an A event that is consistent with the name of the dish that the user P had eaten the last time or the last time before on the base day.

[0070] In addition, the method by which the recommendation unit 7 searches for dish candidates proposed to the user is not limited to collaborative filtering. For example, based on the characteristic index 17 representing the eating habits corresponding to the specific taste sensation information "light" input by the user, statistical analysis, machine learning, or deep learning may be used to search for dish candidates proposed to the user from analysis data 21a1, 21b1, ... containing the eating habits of other users corresponding to the taste sensation information "light".

[0071] For example, Fig.11As shown, when the dishes of the B event of the variables X2, Y1 and Z1 are "chicken stew", "potato stew" and "pork belly cabbage", the recommendation unit 7 can output these dishes as the dish candidates 25a~25c. In addition, the recommendation unit 7 can also output the dish "Nanban pickled fish" of the A event of the variables X2, Y1 and Z1 as the reference dishes 26a~26c. At this time, the recommendation unit 7 can give priority to the variables X2, Y1 and Z1 that record "Nanban pickled fish" as the dish of the A event when the dish that the user P ate the last time before the reference day (for example, 2 days ago) was "Nanban pickled fish". In addition, the recommendation unit 7 can also output the condition-related information 27 related to the selection condition when the variables X2, Y1 and Z1 are selected. Thus, the proposed data 28 including the dish candidates 25a~25c, the reference dishes 26a~26c and the condition-related information 27 is prepared.

[0072] In this way, the recommendation unit 7 outputs the dish candidates 25a to 25c proposed to the user P based on the characteristic indexes 17 such as support, confidence and lift. Here, the characteristic index 17 is a characteristic index obtained by analyzing the characteristics of the eating habits of the user P based on the specific event data 15a corresponding to the specific dish category information input by the user P among the multiple event data 15a, 15b, ..., which are multiple event data in which the identification information, dish category information and eating time information of the dishes that the user P has eaten are classified according to the dish category information in the extraction unit 6. In this way, the recommendation unit 7 can propose a suitable dish to the user P.

[0073] In addition, the recommendation unit 7 analyzes the characteristics of the eating habits of other users P1, P2, ..., and searches for dish candidates 25a to 25c based on the characteristic index 17 of the user P and the characteristic index 19 of other users P1, P2, .... In this way, the recommendation unit 7 outputs dish candidates 25a to 25c based on the eating habits of other users P1, P2, .... Therefore, the recommendation unit 7 can output various dish candidates 25a to 25c other than the eating history of the user P, and can prevent the situation where dishes that the user P has eaten before are proposed multiple times.

[0074] Here, the recommendation unit 7 may also output other information besides the dish candidates 25a to 25c. For example, the recommendation unit 7 may also output the eating habit data representing the temporal relationship between the event data 15z1 and the event data 15z2 based on the analysis data 16a of the user P outputted from the extraction unit 6, and the event data 15z1 and the event data 15z2 are event data including the identification information and the dish category information. In addition, the recommendation unit 7 may also output the eating habit data representing the temporal relationship between the event data 20z1 and the event data 20z2 based on the analysis data 21a1 and 21b1 of other users P1 and P2, and the event data 20z1 and the event data 20z2 are event data including the identification information and the dish category information. For example, Fig.12 As shown, the recommendation unit 7 can also output the eating habit data 32 based on the analysis data 16a of the user P, which represents the temporal relationship between the event data 15z1 and the event data 15z2 with nodes 30 and edges 31, and the event data 15z1 and the event data 15z2 are event data containing the name (identification information) of the dish and the food information. Here, for example, the higher the frequency or support of the event data 15z2 of the event B is selected, the larger the node 30 is formed. In addition, the greater the lift or confidence, the thicker the edge 31 is formed. At this time, the recommendation unit 7 can also limit the specified number of edges 31 to produce the eating habit data 32, and the specified number of edges 31 is the specified number of edges that are ranked in front when sorted from large to small by lift or confidence (for example, limited to the 26 edges 31 that are ranked in front). Furthermore, the recommendation unit 7 may select a path 33 in which the nodes 30 are sequentially connected in order of size for a predetermined number of eating times (eg, 4 times or 4 days), and display the path 33 in a changed display mode.

[0075] In this way, the recommendation unit 7 outputs the eating habit data 32 indicating the temporal relationship between the event data 15z1 and the event data 15z2 based on the characteristic index 17 indicating the characteristics of the eating habit of the user P. In this way, since the unconscious eating habit of the user P is visualized, the user P can easily understand the intention of the dish candidates 25a to 25c in the suggestion data 28 being suggested.

[0076] Next, the recommendation unit 7 sends the created suggestion data 28 to the terminal 1. At this time, the recommendation unit 7 may also send the eating habit data 32 to the terminal 1. Upon receiving the suggestion data 28 and the eating habit data 32, the terminal 1 displays the suggestion data 28 and the eating habit data 32 on the display unit.

[0077] According to the present embodiment, the extraction unit 6 pre-acquires a plurality of event data 15a, 15b, ..., which classify the identification information, the dish category information, and the eating time information of the dishes that the user P has eaten according to the dish category information. When the user P inputs specific dish category information, the extraction unit 6 extracts the characteristic index 17 obtained by analyzing the characteristics of the eating habits of the user P based on the specific event data 15a corresponding to the specific dish category information. Then, the recommendation unit 7 outputs the dish candidates 25a to 25c proposed to the user P based on the characteristic index 17. Thus, the recommendation unit 7 can propose suitable dishes to the user P.

[0078] In addition, in the present embodiment, the recommendation unit 7 outputs the dish candidates 25a to 25c based on the eating habits of other users P1, P2, ..., but is not limited thereto, as long as the dish candidates 25a to 25c can be output. For example, the recommendation unit 7 may select the dish candidates 25a to 25c from the analysis data 16a representing the eating habits of the user P based on the characteristic index 17. For example, the recommendation unit 7 may select the dish candidates 25a to 25c in the analysis data 16a in descending order of the value of the characteristic index 17.

[0079] In addition, in the present embodiment, the dish recommendation system 3 is configured on the server 2, but the present invention is not limited thereto, as long as it can output dish candidates. For example, the dish recommendation system 3 can also be configured on the terminal 1.

[0080] In addition, in the present embodiment, the present disclosure is described using the names of dishes or ingredients of Japanese food, but is not limited thereto. For example, the names of dishes or ingredients may be changed in accordance with the food culture of the country in which the dish recommendation system, dish recommendation method, and program are used.

[0081] Although specific examples of the present disclosure have been described in detail above, these specific examples are merely illustrative and do not limit the claims. The technology described in the claims includes various modifications and changes made to the specific examples described above.

[0082] This application is based on the Japanese patent application (Japanese Patent Application No. 2023-051620) filed on March 28, 2023, the contents of which are incorporated herein by reference.

[0083] Industrial Applicability

[0084] The dish recommendation system disclosed in the present invention can be used in a system that proposes dish candidates to users.

[0085] Description of Reference Numerals

[0086] 1 Terminal

[0087] 2 Server

[0088] 3. Food Recommendation System

[0089] 4. Acquisition

[0090] 5 Storage

[0091] 6 Extraction Department

[0092] 7 Recommended

[0093] 11 Storage Devices

[0094] 12 Processor

[0095] 13 UI Devices

[0096] 14 communication devices

[0097] 15a, 15b Event data

[0098] 16, 16a Analyzing data

[0099] 17 Characteristics

[0100] 18a, 18b Event data

[0101] 21a, 21b Analyze data

[0102] 22 Comparative Data

[0103] 23 Variable Values

[0104] 24 Variable Value

[0105] 25a~25c Dish Candidates

[0106] 26a~26c Reference dishes

[0107] 27 Conditions related information

[0108] 28 Proposed data

[0109] 30 nodes

[0110] 31 Edge

[0111] 32 Eating habits data

[0112] 33 Path

[0113] P User

[0114] P1, P2 Other users

Claims

1. A dish recommendation system, characterized in that: include: an acquisition unit, which pre-acquires a plurality of event data in which identification information, dish category information, and eating time information of dishes eaten by a user are classified according to the dish category information, wherein the dish category information includes taste sensation information in which the dishes are classified according to taste sensation; an extraction unit, which calculates a plurality of characteristic indicators through time series association analysis based on the plurality of event data, and extracts a specific characteristic indicator obtained by analyzing the user's eating habits from the plurality of characteristic indicators based on the specific event data corresponding to the specific dish category information input by the user, wherein the plurality of characteristic indicators are characteristic indicators respectively calculated for each dish category information and include at least one of support, confidence, and lift related to the user's eating habits; and The recommendation unit outputs menu candidates to be proposed to the user based on the specific characteristic index.

2. The dish recommendation system according to claim 1, wherein: The taste sensation information includes information indicating onomatopoeia and mimetic words, which are onomatopoeia and mimetic words that express the sensation of taste.

3. The dish recommendation system according to claim 1, wherein: The plurality of event data include ingredient information of dishes eaten by the user.

4. The dish recommendation system according to claim 3, wherein: The plurality of event data are data formed by associating at least one of the identification information of the dish and the ingredient information with the dish category information.

5. The dish recommendation system according to claim 1, wherein: The extraction unit further extracts the specific characteristic index using at least one of the user's preference information and the avoided food information.

6. The dish recommendation system according to claim 1, wherein: The recommendation unit analyzes the characteristics of the eating habits of other users and searches for the menu candidates based on the specific characteristic index of the user and the characteristic indexes of the other users.

7. The dish recommendation system according to claim 1, wherein: The recommendation unit further outputs the eating habit data representing the temporal relationship of the specific event data using nodes and edges based on the specific characteristic index.

8. A dish recommendation method, characterized in that: The following steps are involved: Acquire in advance a plurality of event data in which identification information of dishes eaten by a user, dish category information, and eating time information are classified according to the dish category information, wherein the dish category information includes taste sensation information in which the dishes are classified according to taste sensation; Based on the plurality of event data, a plurality of characteristic indicators are calculated through time series association analysis, and based on the specific event data corresponding to the specific dish category information input by the user, a specific characteristic indicator obtained by analyzing the user's eating habits is extracted from the plurality of characteristic indicators, wherein the plurality of characteristic indicators are characteristic indicators respectively calculated for each dish category information and include at least one of support, confidence, and lift related to the user's eating habits; and Based on the specific characteristic index, menu candidates proposed to the user are output.

9. A program, characterized in that Causes the computer to perform the following processing: Acquire in advance a plurality of event data in which identification information of dishes eaten by a user, dish category information, and eating time information are classified according to the dish category information, wherein the dish category information includes taste sensation information in which the dishes are classified according to taste sensation; Based on the multiple event data, multiple characteristic indicators are calculated through time series association analysis, and based on the specific event data corresponding to the specific dish category information input by the user, a specific characteristic indicator obtained by analyzing the user's eating habits is extracted from the multiple characteristic indicators, wherein the multiple characteristic indicators are characteristic indicators respectively calculated for each dish category information and include at least one of support, confidence, and lift related to the user's eating habits; as well as Based on the specific characteristic index, menu candidates proposed to the user are output.

Images