Dish proposal device, dish proposal method, dish proposal program, and information processing device

Abstract

The present invention addresses the problem of providing, for example, a dish proposal device capable of proposing dishes that a user may wish to eat, that is, dishes that match the user's preferences (which may include dishes that the user has never eaten and does not know about). A dish proposal device according to the present embodiment comprises: an analyzing unit that uses a clustering model constructed using user data groups relating to dishes, with a plurality of preference characteristics as explanatory variables, and user data of a target user with respect to the dishes, to analyze the target user's characteristic values for each of the plurality of preference characteristics; and a proposing unit that uses the target user's characteristic values for each of the plurality of preference characteristics analyzed by the analyzing unit and the characteristic values of dishes, determined on the basis of information included in a recipe data group, with respect to each of the plurality of preference characteristics associated with each item of recipe data, to propose a dish matching the target user from within a dish group corresponding to the recipe data group.

Description

Recipe suggestion device, recipe suggestion method, recipe suggestion program, and information processing device

[0001] The present invention relates to a recipe suggestion device, a recipe suggestion method, a recipe suggestion program, and an information processing device.

[0002] Patent Literature 1 discloses a recipe information providing device for presenting recipe information suited to a user's preferences. Patent Literature 2 discloses an information processing system that allows a user to input food-specific evaluations. Patent Literature 3 discloses a recommendation presentation device that learns a user's preferences with simple operations and presents recommendations. Patent Literature 4 discloses a computer system that suggests wines to drink with food.

[0003] JP 2019-087064 A JP 2023-113299 A JP 2020-119346 A International Publication No. 2019 / 044341

[0004] However, there is room for improvement in terms of suggesting dishes that users would want to eat, i.e., dishes that suit the user's tastes.

[0005] The present invention has been made in consideration of the above-mentioned problems, and aims to provide a dish suggestion device, a dish suggestion method, and a dish suggestion program that can suggest dishes that a user will want to eat, that is, dishes that suit the user's tastes (which may include dishes that the user has never tried), as well as an information processing device that can contribute to the realization of such suggestions.

[0006] In order to solve the above-mentioned problems and achieve the objectives, the dish suggestion device of the present invention comprises: an analysis unit that analyzes the characteristic values ​​of the target user for each of the plurality of preference characteristics using a clustering model constructed using a group of user data for dishes and a group of preference characteristics as explanatory variables, and the target user's user data for dishes; and a suggestion unit that suggests dishes suitable for the target user from a group of dishes corresponding to the recipe data group, using the characteristic values ​​of the target user for each of the plurality of preference characteristics analyzed by the analysis unit and the characteristic values ​​of dishes for each of the plurality of preference characteristics determined based on information included in the recipe data group and linked to each recipe data.

[0007] The suggestion unit may (1) calculate the match rate between the characteristic value of the target user and the characteristic value of each dish that constitutes the group of dishes based on cosine similarity, (2) extract a predetermined number of dishes from the group of dishes as suggestion candidates based on the calculated match rate, (3) perform a penalty judgment on the extracted dishes to prevent similar dishes from being repeatedly suggested, and (4) suggest a predetermined number of dishes from the extracted dishes based on the results of the penalty judgment.

[0008] In addition, the analysis unit may analyze the characteristic values ​​of the target user while training the clustering model using feedback information of the target user regarding the dish selected by the target user and the current characteristic values ​​of the target user.

[0009] The plurality of preference characteristics may also be composed of a plurality of preference characteristics belonging to the category of taste, a plurality of preference characteristics belonging to the category of texture, or a plurality of preference characteristics belonging to the category of flavor or aroma.

[0010] Furthermore, the plurality of preference characteristics may be composed of a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of texture, a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of flavor or aroma, or a combination of one or more preference characteristics belonging to the category of texture with one or more preference characteristics belonging to the category of flavor or aroma.

[0011] The plurality of preference characteristics may also be composed of a combination of one or more preference characteristics belonging to the category of taste, one or more preference characteristics belonging to the category of texture, and one or more preference characteristics belonging to the category of flavor or aroma.

[0012] In addition, the dish suggestion device according to the present invention may further include an annotation unit that determines a characteristic value of a dish for each of the plurality of preference characteristics based on the information contained in the recipe data group and links the determined characteristic value to each recipe data.

[0013] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least one of the texts included in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0014] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least two of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0015] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least three of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0016] In addition, the annotation unit may determine characteristic values ​​for each dish for each of the multiple preference characteristics based on text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantities, and text regarding the cooking process contained in the recipe data.

[0017] The annotation unit may also determine the characteristic value of a dish by assigning, to each of the plurality of preference characteristics, either a value indicating that the characteristic applies or a value indicating that the characteristic does not apply.

[0018] In addition, the dish suggestion method of the present invention includes an analysis step in which an analysis unit analyzes the characteristic values ​​of the target user for each of the multiple preference characteristics using a clustering model constructed using a group of user data for dishes with multiple preference characteristics as explanatory variables, and the user data for the target user for dishes; and a suggestion step in which a suggestion unit suggests a dish that suits the target user from a group of dishes corresponding to the recipe data group using the characteristic values ​​of the target user for each of the multiple preference characteristics analyzed in the analysis step and the characteristic values ​​of dishes for each of the multiple preference characteristics that are determined based on information included in the recipe data group and linked to each recipe data.

[0019] In addition, the dish suggestion program of the present invention causes a computer to function as: an analysis means for analyzing the characteristic values ​​of the target user for each of the plurality of preference characteristics using a clustering model constructed using a group of user data for dishes with a plurality of preference characteristics as explanatory variables, and the user data for the target user for dishes; and a suggestion means for suggesting a dish suitable for the target user from a group of dishes corresponding to the recipe data group using the characteristic values ​​of the target user for each of the plurality of preference characteristics analyzed by the analysis means, and the characteristic values ​​of dishes for each of the plurality of preference characteristics determined based on information contained in the recipe data group and linked to each recipe data.

[0020] The recording medium according to the present invention is a non-transitory computer-readable recording medium having the recipe suggestion program recorded thereon, i.e., a non-transitory computer-readable recording medium including programmed instructions for causing a computer to execute the recipe suggestion method.

[0021] The information processing device according to the present invention also includes an annotation unit that determines a characteristic value of a dish for each of a plurality of preference characteristics based on information included in a recipe data group and associates the determined characteristic value with each recipe data.

[0022] The plurality of preference characteristics may be composed of a plurality of preference characteristics belonging to the category of taste, a plurality of preference characteristics belonging to the category of texture, or a plurality of preference characteristics belonging to the category of flavor or aroma.

[0023] Furthermore, the plurality of preference characteristics may be composed of a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of texture, a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of flavor or aroma, or a combination of one or more preference characteristics belonging to the category of texture with one or more preference characteristics belonging to the category of flavor or aroma.

[0024] The plurality of preference characteristics may also be composed of a combination of one or more preference characteristics belonging to the category of taste, one or more preference characteristics belonging to the category of texture, and one or more preference characteristics belonging to the category of flavor or aroma.

[0025] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least one of the texts included in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0026] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least two of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0027] In addition, the annotation unit may determine the characteristic values ​​of the dish for each of the multiple preference characteristics based on at least three of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

[0028] In addition, the annotation unit may determine characteristic values ​​for each dish for each of the multiple preference characteristics based on text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantities, and text regarding the cooking process contained in the recipe data.

[0029] The present invention has the effect of being able to suggest dishes that the user will want to eat, that is, dishes that suit the user's tastes (which may include dishes that the user has never tried), and also has the effect of contributing to the realization of the suggestions.

[0030] FIG. 1 is a diagram showing an example of the configuration of the recipe recommendation device 100. FIG. 2 is a diagram showing an overview of various processes executed by the recipe recommendation device 100. FIG. 3 is a diagram showing an example of an annotation implementation method. FIG. 4 is a diagram showing an example of annotation rules. FIG. 5 is a diagram showing an example of annotation rules. FIG. 6 is a diagram showing an example of a rule pattern. FIG. 7 is a diagram showing an example of the results of clustering each dish using coordinate data, which is obtained by performing correspondence analysis on the characteristic values ​​of several thousand to nearly 10,000 dishes registered in a recipe database for each of a plurality of predefined preference characteristic tags. FIG. 8 is a diagram showing an overview of the analysis process executed by the recipe recommendation device 100. FIG. 9 is a diagram showing an overview of the analysis process and recommendation process executed by the recipe recommendation device 100.

[0031] Hereinafter, embodiments of a recipe recommendation device, a recipe recommendation method, a recipe recommendation program, and an information processing device according to the present invention will be described in detail with reference to the accompanying drawings. However, the present invention is not limited to the embodiments.

[0032] 1 is a diagram showing an example of the configuration of a recipe recommendation device 100. The recipe recommendation device 100 is communicably connected to, for example, an information communication terminal 200 (e.g., a mobile phone, smartphone, tablet terminal, personal computer, etc.) owned by a user (ordinary consumer) via a network 300 such as the Internet, an intranet, or a LAN (including both wired and wireless networks). The recipe recommendation device 100 may also be communicably connected to a server device (not shown).

[0033] The recipe recommendation device 100 includes a control unit 102 that comprehensively controls devices such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), a communication interface unit 104 that communicatively connects the device to a network 300 via a communication device such as a router and a wired or wireless communication line such as a dedicated line, a storage unit 106 that stores various databases, tables, files, etc., and an input / output interface unit 108 that connects to an input device 112 and an output device 114. The various units included in the recipe recommendation device 100 are communicatively connected via any communication path.

[0034] The communication interface unit 104 mediates communication between the dish recommendation device 100 and the network 300 (or a communication device such as a router). In other words, the communication interface unit 104 has a function of communicating data with other terminals via a communication line.

[0035] An input device 112 and an output device 114 are connected to the input / output interface unit 108. The output device 114 may be a monitor, a speaker, a printer, or the like. The input device 112 may be a keyboard, a mouse, a microphone, a monitor that functions as a pointing device in cooperation with a mouse, or a touch panel. Information such as user preference characteristics and recipe data information may be input from images. A camera or the like may also be used as the input device 112.

[0036] The memory unit 106 is a storage means. For example, a memory device such as a RAM or ROM, a fixed disk device such as a hard disk, a flexible disk, or an optical disk can be used as the memory unit 106. The memory unit 106 may store a computer program that works in cooperation with an operating system (OS) to issue commands to the CPU or GPU to perform various processes.

[0037] The storage unit 106 includes a recipe database 106a related to recipes for single dishes. If a server device (not shown) is communicatively connected to the dish suggestion device 100, the recipe database 106a may be stored in the server device.

[0038] The control unit 102 has an internal memory for storing control programs such as an OS, programs that define various processing procedures, required data, etc., and executes various information processing based on these programs. Functionally, the control unit 102 includes an annotation unit 102a, an analysis unit 102b, and a proposal unit 102c.

[0039] The annotation unit 102a determines a characteristic value of each dish for each of a plurality of predefined preference characteristic tags based on the information contained in the recipe database 106a and in accordance with a predetermined annotation rule, and associates the determined characteristic value with each recipe data (see FIGS. 2-7). The preference characteristic tag corresponds to the preference characteristic of the present invention. FIG. 2 shows an overview of the annotation process executed by the annotation unit 102a. FIG. 3 shows an example of an annotation implementation method. FIGS. 4 and 5 show examples of annotation rules. FIG. 6 shows an example of a rule pattern included in the annotation rule. FIG. 7 shows an example of the results of performing correspondence analysis on the characteristic values ​​(categorical data of 0 or 1) of several thousand to nearly 10,000 dishes registered in the recipe database 106a for each of a plurality of predefined preference characteristic tags, thereby representing each dish as coordinate data, and then clustering the dishes using this coordinate data using a predetermined clustering method (e.g., k-means). The control unit 102 may further include a processing unit (not shown) that performs clustering of dishes. Furthermore, the results of the clustering of dishes may be used, for example, to set multiple dishes to be presented when a new user is asked to answer questions about their likes and dislikes of dishes. Specifically, multiple dishes belonging to different clusters may be set as the multiple dishes to be presented.

[0040] The annotation unit 102a may determine the characteristic value of each of a plurality of predefined preference characteristics of a dish based on at least one of the following texts included in the recipe data: text about the name of the dish, text about the names of ingredients and seasonings, text about the amounts, and text about the cooking steps (see the descriptions about the recipe and cooking method in FIGS. 4 and 5 ) (see FIG. 3 ). The annotation unit 102a preferably uses at least two of these four texts, more preferably uses at least three of these four texts, and more preferably uses all four texts. The annotation unit 102a may also employ a concept such as a "penalty" (described below). For example, a description of a seasoning, such as "preferred amount," may not be used as annotation text information.Text related to cooking processes includes, for example, "stir-fry," "deep-fry," "boil," "steam," "bake," "pickle," "mix," "cook," "fry," "boil," "cover," "steam," "bake," "cook in a pan," "sear," "broil," "knead," "color," "blanch," "steam," "finish," "simmer," "stew," "bake," "mix," "mix," "strain," "chop," "rub," "shave," "cut," "wrap," "roll," "shape," "pound," "shake," "melt," "knead," "press," "warm," "cool," "add flavor," "season," "brown," "sauté," "roast," "marinate," "grill," "bread with breadcrumbs," "sprinkle with Parmesan cheese," "fry," "butter," "make into a gratin," "sear," "make a roux," "garnish," "tear," "soak," "grab," "arrange," "dry," "make," "use," "put in," "cool," "divide," "cut," "break" "," "add," "move," "take out," "remove," "take," "combine," "contain," "turn," "solidify," "fill," "paint," "remove," "gather," "lay," "roll," "spread," "pull up," "pull out," "weaken," "shape," "put back," "fold," "press," "wipe," "hold," "pinch," "discard," "grasp," "let go," "scatter," "arrange," "lay down," "cut," "write," "stop," "leave," "boil," "whisk," "pouring," "wash," "drain," "soak," "add," "warm," "moisten," "dissolve," "soak," "cook," "boil out," "boil," "thaw," "heat," "plat," "crush," "transfer," "stand," "stand," "combine," "squeeze," "knead," "sew," "repeat," "drop," "tear," "supplement," "cover," "pack," "add," "roll," "return," "send," "escape," "cook," "select," "layer," "close," "open," "remove," "fly," or "decorate."

[0041] In addition, the annotation unit 102a may determine the characteristic value of a dish by assigning either a value indicating that the dish applies (e.g., 1) or a value indicating that the dish does not apply (e.g., 0) to each of a plurality of predefined preference characteristics.

[0042] Here, the predefined plurality of preference characteristic tags may be composed of, for example, (1) a plurality of preference characteristic tags selected from a set of preference characteristic tags in the category of taste (e.g., sweet, sour, bitter, spicy, etc.), (2) a plurality of preference characteristic tags selected from a set of preference characteristic tags in the category of texture (e.g., crunchy, fluffy, soft, chewy, etc.), or (3) a plurality of preference characteristic tags selected from a set of preference characteristic tags in the category of flavor or aroma (e.g., bonito broth flavor, butter flavor, nutmeg flavor, ethnic flavor, etc.).

[0043] Furthermore, the predefined plurality of preference characteristic tags may be configured, for example, as follows: (1) a combination of one or more preference characteristic tags selected from the set of preference characteristic tags in the category of taste and one or more preference characteristic tags selected from the set of preference characteristic tags in the category of texture; (2) a combination of one or more preference characteristic tags selected from the set of preference characteristic tags in the category of taste and one or more preference characteristic tags selected from the set of preference characteristic tags in the category of flavor or aroma; or (3) a combination of one or more preference characteristic tags selected from the set of preference characteristic tags in the category of texture and one or more preference characteristic tags selected from the set of preference characteristic tags in the category of flavor or aroma.

[0044] In addition, the multiple predefined preference characteristic tags may be composed of, for example, a combination of one or more preference characteristic tags selected from a set of preference characteristic tags in the category of taste, one or more preference characteristic tags selected from a set of preference characteristic tags in the category of texture, and one or more preference characteristic tags selected from a set of preference characteristic tags in the category of flavor or aroma (see Figures 3-5).

[0045] Returning to FIG. 1 , the analysis unit 102b uses "a clustering model constructed in advance by a predetermined clustering method (e.g., k-means) using a group of user data for dishes (e.g., single dishes) and a plurality of predefined preference characteristic tags as explanatory variables" and "user data for the target user regarding dishes (e.g., data related to initial questions, behavioral data, etc.)" to analyze the characteristic values ​​of the target user for each of the plurality of preference characteristic tags (see FIGS. 2 and 8-9). In other words, the analysis unit 102b identifies the preferences of the target user using the preference characteristic tags. FIG. 2 shows an overview of the analysis process performed by the analysis unit 102b. FIG. 8 shows an overview of the analysis process performed by the analysis unit 102b upon initial registration. FIG. 9 shows an overview of the analysis process performed by the analysis unit 102b upon continued use after initial registration.

[0046] Note that, at the time of initial registration, the analysis unit 102b may use a clustering model to identify clusters that match the target user, thereby analyzing (inferring) the characteristic values ​​of the target user at the time of initial registration (see Figures 2 and 8). Specifically, the analysis unit 102b may execute the processes shown in (11) to (13) below. (11) Using the target user's current characteristic values ​​and feedback information on the dishes selected by the target user, a weighted activation function (e.g., a softmax function) is used to calculate the characteristic values ​​of the target user for continued use for each preference characteristic tag. (12) Using the calculated coordinate data, a predetermined clustering method (e.g., k-means method, etc.) is used to construct multiple clusters (e.g., 10 to 20 clusters). Note that when the k-means method is used as the clustering method, the number of clusters may be determined using existing methods such as the elbow method or the silhouette coefficient. (13) Cluster classification is performed using a clustering model using the arithmetic mean of the characteristic values ​​of each preference characteristic tag, taking into account the likes and dislikes of multiple types of cuisine and their weighting. The center point of the classified cluster is set as the initial characteristic value of the user. For cluster classification, the arithmetic mean of the "characteristic value of cuisine" x "feedback_rate (e.g., like: 0.5, dislike: -0.5)" is input to the clustering model. As shown in Figure 8, even with a small number of questions (four cuisines x four questions), the cluster to which the user belongs can be determined with a probability of approximately 26% (compared to an 8% probability with random determination). Furthermore, the probability of recommending the user's preferred cuisine was high at approximately 67%, confirming that a similar cluster can be selected even if the initial cluster determination was incorrect.

[0047] Furthermore, when the target user continues to use the service after the initial registration, the analysis unit 102b may analyze (update) the characteristic values ​​of the target user while training a clustering model using feedback information about the target user regarding the dishes selected by the target user (e.g., feedback information obtained through a service such as menu suggestions) and the target user's current characteristic values ​​(see FIGS. 2 and 9). As a result, even between users who belong to the same cluster at the time of initial registration, the preferences are updated to reflect each user's individuality based on the feedback information (e.g., answer selection of dishes). Specifically, the analysis unit 102b may use the target user's current characteristic values ​​and the feedback information about the dishes selected by the target user to multiply the arithmetic mean of the characteristic values ​​of each preference characteristic tag, taking into account the target user's likes and dislikes and their weighting, by the target user's current characteristic value and an activation function (e.g., an inverse sigmoid function) with tuned slope and bias, and add the result to the target user's current characteristic value, thereby calculating the characteristic value for each preference characteristic tag when the target user continues to use the service.

[0048] Returning to FIG. 1 , the suggestion unit 102c uses the characteristic values ​​of the target user for each of the predefined preference characteristic tags analyzed by the analysis unit 102b and the characteristic values ​​of the dishes for each of the predefined preference characteristic tags determined and linked to each recipe data by the annotation unit 102a to suggest dishes suitable for the target user from among the group of dishes corresponding to the recipe database 106a (i.e., the group of dishes registered in the recipe database 106a) (see FIGS. 2 and 9 ). That is, the suggestion unit 102c suggests dishes that match the target user based on the characteristic values ​​of the target user learned by the analysis unit 102b and the characteristic values ​​of the recipes registered in the annotation unit 102a. That is, the suggestion unit 102c suggests appropriate dishes to the target user based on the target user's preferences. FIGS. 2 and 9 show an overview of the suggestion process performed by the suggestion unit 102c.

[0049] The suggestion unit 102c may execute the following processes (1) to (4): (1) Calculate (score) the match rate between the characteristic value of the target user (specifically, a vector having the number of defined preference characteristic tags as its dimension and the characteristic value of each preference characteristic tag as its component) and the characteristic value of each dish constituting the group of dishes corresponding to the recipe database 106a (specifically, a vector having the number of defined preference characteristic tags as its dimension and the characteristic value of each preference characteristic tag as its component) based on cosine similarity. (2) Based on the calculated match rate (score), extract a predetermined number of dishes from the group of dishes as suggestion candidates. (3) For the dishes extracted in (2), execute a penalty determination based on the recipe data to prevent similar dishes from being repeatedly suggested. ※１Ａ※１Ｂ *1A: For example, the extracted dishes are grouped based on the condition that the recipe is an exact match or the cooking method, ingredients, or genre are the same, and as a result, if there is a group consisting of multiple dishes, a penalty is imposed to prevent any of the dishes in the group from being repeatedly suggested. *1B: Similar dishes may be grouped using language processing, and a penalty judgment may be made to "not suggest dishes from the same group consecutively." (4) Based on the result of the penalty judgment, a predetermined number of dishes are suggested from the dishes extracted in (2). ※２※３ *2: The dish may be suggested as is. *3: The dish may be suggested based on serendipity. For example, a dish may be suggested that has the same recipe as the dish selected based on the judgment results, but with some differences in the ingredients used (such as ingredients or seasonings). For example, the dishes may be clustered, and dishes may be suggested that belong to a cluster close to the cluster to which the dish selected based on the judgment results belongs.

[0050] As described above, according to this embodiment, it is possible to suggest dishes that the user wants to eat, that is, dishes that suit the user's preferences (which may include dishes that the user has never tried). According to this embodiment, dishes are suggested from the same cluster, so dishes that suit the user's preferences are suggested regardless of whether the user has eaten them before. Therefore, dishes that use ingredients or seasonings that the user has never tried before may also be suggested. Therefore, this embodiment has the potential to provide the user with the surprising and delightful experience of accidentally encountering something unexpected. In other words, this embodiment can create serendipity (accidental and fortunate discoveries) for the user.

[0051] Here, existing food recommendation technologies have problems such as not being able to suggest recipes that suit a user's preferences unless the user uses the system extensively, tending to suggest similar menus, and making it difficult to find a dish that suits a user's preferences from the recommendations. Therefore, by adopting the above-described configuration and processing, this embodiment makes it possible to suggest recipes that suit a user's preferences, and even to suggest recipes that suit a user's preferences even if the user is unfamiliar with the dish.

[0052] The annotation technology described above can be used in the following ways (1) to (9), starting from this technology. Furthermore, statistical processing and machine learning (such as supervised learning or reinforcement learning) can also be used for the annotation and recipe suggestion technologies described above. (1) Annotation technology can be used alone to evaluate and analyze market trends, the characteristics of one's own and other companies' products and recipes, eating habits, and food culture. (2) Annotation technology can be combined with preference analysis technology to provide a personal preference analysis service. (3) Annotation technology can be combined with preference analysis technology and recommendation technology to provide a service (e.g., food and recipe suggestions) based on personal preference analysis. (4) Annotation technology can be combined with preference analysis technology and recipe generation technology to generate recipes tailored to individual preferences. (5) Annotation technology can be combined with preference analysis technology and cooking technology using cooking appliances / robots to automatically prepare meals tailored to individual preferences. (6) Annotation technology can be used to generate advertising expressions by combining it with large-scale language models. (7) Annotation technology can be used to develop products by combining it with POS data. (8) Annotation technology can be used to develop products by combining it with preference analysis, recommendation technology, and location information acquisition technology. (9) Annotation technology can be used to improve the presentation method for recipe searches by combining it with search technology.

[0053] [2. Other Embodiments] While the embodiments of the present invention have been described above, the present invention may be embodied in various different embodiments other than those described above within the scope of the technical concept set forth in the claims.

[0054] For example, among the processes described in the embodiments, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically using known methods.

[0055] In addition, the processing procedures, control procedures, specific names, registered data for each process, information including parameters such as search conditions, screen examples, and database configurations shown in the above documents and drawings can be changed as desired unless otherwise specified.

[0056] Furthermore, with regard to each device, the components shown in the drawings are functional concepts, and do not necessarily have to be physically configured as shown in the drawings.

[0057] For example, all or any part of the processing functions of the recipe recommendation device 100, particularly those performed by the control unit 102, may be implemented by a CPU or GPU and a program interpreted and executed by the CPU or GPU, or may be implemented as hardware using wired logic. The program is recorded on a non-transitory computer-readable recording medium containing programmed instructions for causing the information processing device to execute the processes described in the embodiments, and is mechanically read by the recipe recommendation device 100 as needed. That is, the storage unit 106, such as a ROM or HDD, stores a computer program that works in conjunction with the OS to issue commands to the CPU or GPU and perform various processes. The computer program is executed by being loaded into RAM and cooperates with the CPU or GPU to form the control unit.

[0058] In addition, this computer program may be stored in an application program server connected to the dish recommendation device 100 via any network, and all or part of it may be downloaded as needed.

[0059] The recipe recommendation program of the present invention may be stored in a non-transitory computer-readable recording medium or configured as a program product. Here, the term "recording medium" includes any "portable physical medium" such as a memory card, USB memory, SD card, flexible disk, magneto-optical disk, ROM, EPROM, EEPROM, CD-ROM, MO, DVD, and Blu-ray (registered trademark) Disc.

[0060] Furthermore, a "program" is a data processing method written in any language or description method, regardless of the format, such as source code or binary code. Note that a "program" is not necessarily limited to a single program, but also includes programs that are distributed as multiple modules or libraries, or programs that achieve their functions by working together with other programs, such as an OS. Note that the specific configurations and reading procedures for reading a recording medium in each device shown in the embodiments, as well as the installation procedures after reading, can use well-known configurations and procedures.

[0061] The various databases stored in the memory unit 106 are storage means such as memory devices such as RAM and ROM, fixed disk devices such as hard disks, flexible disks, and optical disks, and store various programs, tables, databases, and web page files used for various processes and providing websites.

[0062] The recipe suggestion device 100 may be configured as an information processing device such as a known personal computer or workstation, or may be configured as an information processing device connected to any peripheral device. The recipe suggestion device 100 may also be realized by installing software (including programs, data, etc.) that causes the information processing device to implement the recipe suggestion method of the present invention.

[0063] Furthermore, the specific form of distribution and integration of the devices is not limited to that shown in the drawings, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various additions or functional loads. In other words, the above-mentioned embodiments can be implemented in any combination, or embodiments can be implemented selectively.

[0064] The present invention is extremely useful in many industrial fields, particularly in services that provide cooking recipes and the like over the Internet.

[0065] REFERENCE SIGNS LIST 100 Recipe suggestion device 102 Control unit 102a Annotation unit 102b Analysis unit 102c Proposal unit 104 Communication interface unit 106 Storage unit 106a Recipe database 108 Input / output interface unit 112 Input device 114 Output device 200 Information communication terminal 300 Network

Claims

1. A cooking suggestion device comprising: an analysis unit that analyzes a characteristic value of a target user for each of the plurality of preference characteristics using a clustering model constructed using a group of user data for dishes with a plurality of preference characteristics as explanatory variables, and the user data for the target user for the dishes; and a suggestion unit that suggests a dish that suits the target user from a group of dishes corresponding to the recipe data group, using the characteristic value of the target user for each of the plurality of preference characteristics analyzed by the analysis unit and the characteristic value of a dish for each of the plurality of preference characteristics that is determined based on information included in the recipe data group and linked to each recipe data.

2. The dish suggestion device of claim 1, wherein the suggestion unit (1) calculates the match rate between the characteristic value of the target user and the characteristic value of each dish that constitutes the group of dishes based on cosine similarity, (2) extracts a predetermined number of dishes from the group of dishes as suggestion candidates based on the calculated match rate, (3) performs a penalty judgment on the extracted dishes to prevent similar dishes from being repeatedly suggested, and (4) suggests a predetermined number of dishes from the extracted dishes based on the results of the penalty judgment.

3. The cooking suggestion device according to claim 1 or 2, wherein the analysis unit analyzes the characteristic values ​​of the target user while training the clustering model using feedback information of the target user regarding the dish selected by the target user and the current characteristic values ​​of the target user.

4. The cooking suggestion device of claim 3, wherein the plurality of preference characteristics are composed of a plurality of preference characteristics belonging to the category of taste, a plurality of preference characteristics belonging to the category of texture, or a plurality of preference characteristics belonging to the category of flavor or aroma.

5. The cooking suggestion device of claim 4, wherein the plurality of preference characteristics are composed of a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of texture, a combination of one or more preference characteristics belonging to the category of taste with one or more preference characteristics belonging to the category of flavor or aroma, or a combination of one or more preference characteristics belonging to the category of texture with one or more preference characteristics belonging to the category of flavor or aroma.

6. The cooking suggestion device of claim 5, wherein the plurality of preference characteristics are composed of a combination of one or more preference characteristics belonging to the category of taste, one or more preference characteristics belonging to the category of texture, and one or more preference characteristics belonging to the category of flavor or aroma.

7. The cooking suggestion device according to claim 6, further comprising an annotation unit that determines a characteristic value of a dish for each of the plurality of preference characteristics based on the information contained in the recipe data group and links the determined characteristic value to each recipe data.

8. The cooking suggestion device described in claim 7, wherein the annotation unit determines the characteristic value of the dish for each of the plurality of preference characteristics based on at least one of texts included in the recipe data, including text about the name of the dish, text about the names of ingredients and seasonings, text about the amount, and text about the cooking process.

9. The cooking suggestion device described in claim 8, wherein the annotation unit determines the characteristic value of the dish for each of the plurality of preference characteristics based on at least two of the following texts included in the recipe data: text about the name of the dish; text about the names of ingredients and seasonings; text about the quantity; and text about the cooking process.

10. The cooking suggestion device described in claim 9, wherein the annotation unit determines the characteristic values ​​of the dish for each of the plurality of preference characteristics based on at least three of the following texts contained in the recipe data: text about the name of the dish; text about the names of ingredients and seasonings; text about the quantity; and text about the cooking process.

11. The cooking suggestion device described in claim 10, wherein the annotation unit determines characteristic values ​​of each dish for each of the multiple preference characteristics based on text about the name of the dish, text about the names of ingredients and seasonings, text about the amounts, and text about the cooking process contained in the recipe data.

12. The dish suggestion device according to claim 11, wherein the annotation unit determines the characteristic value of the dish by assigning, to each of the plurality of preference characteristics, either a value indicating that the preference characteristics apply or a value indicating that the preference characteristics do not apply.

13. A dish suggestion method comprising: an analysis step in which an analysis unit analyzes the characteristic values ​​of the target user for each of the plurality of preference characteristics using a clustering model constructed using a group of user data for dishes with a plurality of preference characteristics as explanatory variables, and the user data for dishes of the target user; and a suggestion step in which a suggestion unit suggests a dish that suits the target user from a group of dishes corresponding to the recipe data group, using the characteristic values ​​of the target user for each of the plurality of preference characteristics analyzed in the analysis step and the characteristic values ​​of dishes for each of the plurality of preference characteristics that are determined based on information included in the recipe data group and linked to each recipe data.

14. A dish suggestion program that causes a computer to function as: an analysis means that analyzes the characteristic values ​​of the target user for each of the multiple preference characteristics using a clustering model constructed using a group of user data for dishes with multiple preference characteristics as explanatory variables, and the user data for the target user for dishes; and a suggestion means that suggests dishes that suit the target user from a group of dishes corresponding to the recipe data group, using the characteristic values ​​of the target user for each of the multiple preference characteristics analyzed by the analysis means, and the characteristic values ​​of dishes for each of the multiple preference characteristics that are determined based on information included in the recipe data group and linked to each recipe data.

15. An information processing device having an annotation unit that determines a characteristic value of a dish for each of a plurality of preference characteristics based on information contained in a recipe data group and links the value to each recipe data.

16. The information processing device according to claim 15, wherein the plurality of preference characteristics are composed of a plurality of preference characteristics belonging to the category of taste, a plurality of preference characteristics belonging to the category of texture, or a plurality of preference characteristics belonging to the category of flavor or aroma.

17. The information processing device of claim 16, wherein the plurality of preference characteristics are composed of a combination of one or more preference characteristics belonging to the category of taste and one or more preference characteristics belonging to the category of texture, a combination of one or more preference characteristics belonging to the category of taste and one or more preference characteristics belonging to the category of flavor or aroma, or a combination of one or more preference characteristics belonging to the category of texture and one or more preference characteristics belonging to the category of flavor or aroma.

18. The information processing device described in claim 17, wherein the plurality of preference characteristics are composed of a combination of one or more preference characteristics belonging to the category of taste, one or more preference characteristics belonging to the category of texture, and one or more preference characteristics belonging to the category of flavor or aroma.

19. An information processing device described in any one of claims 15 to 18, wherein the annotation unit determines the characteristic value of the dish for each of the plurality of preference characteristics based on at least one of texts included in the recipe data, including text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding quantities, and text regarding the cooking process.

20. The information processing device described in claim 19, wherein the annotation unit determines the characteristic values ​​of the dish for each of the plurality of preference characteristics based on at least two of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

21. The information processing device described in claim 20, wherein the annotation unit determines the characteristic values ​​of the dish for each of the plurality of preference characteristics based on at least three of the following texts contained in the recipe data: text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the quantity, and text regarding the cooking process.

22. The information processing device described in claim 21, wherein the annotation unit determines characteristic values ​​of each dish for each of the multiple preference characteristics based on text regarding the name of the dish, text regarding the names of ingredients and seasonings, text regarding the amounts, and text regarding the cooking process contained in the recipe data.

Images