KR20220148979A - Menu recommendation server, system and method for recommending a personalized menu using it - Google Patents

Abstract

The present invention relates to a diet recommendation server and a personal diet recommendation service system and method, and more specifically to the server for recommending a diet by deep learning of a gated recurrent unit (GRU) algorithm which introduces reset variables and update variables and the system and method for recommending a personal diet using the same. The diet recommendation server according to an embodiment of the present invention has a diet recommendation learning unit using a previous meal of a user as an input layer and taste (diet) or pattern information of the user as a hidden layer to recommend a future diet as an output layer by deep learning previous data with a GRU method. Accordingly, the personal diet can be automatically recommended.

Description

{Menu recommendation server, system and method for recommending a personalized menu using it}

The present invention relates to a meal recommendation server and a personal meal recommendation service system and method, and more particularly, to a server for recommending a meal by deep learning of a GRU (Gated Recurrent Unit) algorithm that introduces a reset variable and an update variable, and a server using the same It relates to a system and method for recommending a personal diet.

Recently, as the quality of life increases, interest in health care tailored to changes in lifestyle and environment is also increasing.

In addition, as more and more people are trying to keep their health in their daily life, not just in hospitals, they are paying great attention to their diet.

It is very difficult for non-professionals to directly plan a healthy diet for themselves with general information found on the Internet or in books, so research on diet recommendation methods that can help diet using various recommendation techniques is being actively conducted. is becoming

However, most of the current diet recommendation systems and diet-related applications on the market focus only on purposes such as diet or health regardless of the individual's physical condition or disease, and output a set diet in turn, not a recommendation considering various aspects. This ensures that all users receive the same diet.

That is, there is a problem in that the recommended diet is set and the user's situation is not taken into account at all, so there is a problem in that it is not possible to perform a customized recommendation by adaptively considering time information including the season, individual dietary information, and surrounding food material information.

Publication No. 10-2009-0046994 (Published on May 12, 2009)

The present invention has been devised to solve the above problems, and it is possible to automatically recommend a diet suitable for an individual from a large amount of data without setting a criterion for judgment using a deep learning algorithm (method). The purpose of the present invention is to provide a personal diet recommendation service system and method using a deep learning algorithm.

In addition, another object of the present invention is to provide a user with a variety of personal circumstances by automatically recommending a diet suitable for an individual using a deep learning algorithm from data such as time information including seasons, individual dietary information, and surrounding food material information. An object of the present invention is to provide a personal diet recommendation service system and method that can recommend a suitable diet.

A meal recommendation server according to an embodiment of the present invention for achieving the above object is a Gated Recurrent Unit (GRU) by using the meal that the user ate immediately before as an input layer and the user's taste (eating) or pattern information as a hidden layer. ) method to deep-learning previous data to provide an output layer with a menu recommendation learning unit that recommends subsequent meals.

In addition, the GRU-type model introduces a reset gate and an update gate to each node of the hidden layer,

The diet recommendation learning unit includes a time variable learning module corresponding to a reset gate that determines how much the past (previous) pattern (diet) affects future patterns (recommended diet) by deep learning, and a new input by deep learning. A reflection learning module corresponding to an update gate that determines how much ingredients and ingredients of the previous diet will be reflected in a future recommended diet, and a recommendation module for recommending a meal based on the reset variable of the time variable learning module and the update variable of the reflection learning module characterized in that it consists of

In addition, the meal recommendation learning unit learns by using menus such as kimchi stew and rolled eggs as one type of input layer, learns by classifying each ingredient instead of a menu, or by classifying meal times such as breakfast/lunch/dinner Or, it is characterized in that it learns including the user's taste (eating) or health information.

And a menu information DB for storing user information, meal information, standard meal information, food recipe, and ingredient information required for the meal recommendation server to recommend a meal, and a matching standard meal matching the meal recommended or recommended by the recommendation module It is characterized in that the display unit for displaying is further provided.

A personal diet recommendation service system according to an embodiment of the present invention, any one of the above-described menu recommendation server;

a user terminal accessing the meal recommendation server, entering personal information to join as a member, and using a personal meal recommendation service provided by the meal recommendation server; and

a shopping mall terminal that connects to the meal recommendation server and inputs food material information of a nearby shopping mall to provide assistance when recommending a meal by the deep learning of the meal recommendation server and sells food materials; is comprised of

In addition, the menu recommendation server displays a recommended diet or a matching standard diet on the screen of the user terminal, and when one of them is selected, a food recipe of the corresponding diet or information of a store selling ingredients of the corresponding diet is displayed. characterized.

The personal diet recommendation service method according to an embodiment of the present invention comprises the steps of (a) determining how much a past (previous) pattern (diet) affects a future pattern (recommended diet) by deep learning in a time variable learning module ;

(b) receiving new ingredient information from the outside and determining how much to reflect the ingredients of the previous diet and newly input ingredients by deep learning in the reflection learning module in the future recommended diet; and

(c) recommending a diet based on the values determined in steps (a) and (b) in the recommendation module;

is comprised of

In addition, before the step (a), it is characterized in that personal information such as personal diet or health information is provided.

According to the means for solving the above-mentioned problems, it is possible to automatically recommend a diet suitable for an individual from a large amount of data without setting a criterion for judgment using a deep learning algorithm (method).

In addition, by using a deep learning algorithm from data such as time information including seasons, individual dietary information, and surrounding food material information, a diet suitable for an individual is automatically recommended. have.

1 is a block diagram of a personal diet recommendation service system according to an embodiment of the present invention.
2 is a diagram showing a GRU model applied to deep learning of the present invention.
3 is a flowchart of a personal diet recommendation service method according to an embodiment of the present invention.
4 is a diagram illustrating a method for implementing a diet recommendation by deep learning of the present invention.
5 is a flowchart illustrating a learning method by deep learning of the present invention.
6 is an exemplary diagram of a method for implementing a diet recommendation according to an embodiment of the present invention.

Hereinafter, the configuration and operation of the embodiment of the present invention will be described with reference to the accompanying drawings.

It should be noted that the same components in the drawings are indicated by the same reference numbers and symbols as much as possible even though they are indicated in different drawings.

In the following description of the present invention, if it is determined that a detailed description of a related well-known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, when a part "includes" a certain component, this means that other components may be further included, rather than excluding other components, unless otherwise stated.

In order to help the understanding of the present invention, deep learning will be described as follows.

Deep learning consists of an input layer that receives and sends an input, a hidden layer that is composed of several nodes and responds accordingly by receiving a stimulus from the input at each node, and It consists of an output layer that sums the responses of the nodes and emits them.

In such deep learning, a recurrent neural network (RNN) is a neural network in which past data can influence the future. It can be seen that it has memory information.

The GRU (Gated Recurrent Unit) is a form of the development of the recurrent neural network, and the influence of the input decreases over time (inevitably, the previous input is forgotten later). Information can be stored for a long period without loss through cells and gates.

2 is a diagram illustrating a GRU model applied to deep learning of the present invention, and a reset gate and an update gate are introduced at each node of the hidden layer.

The reset gate r calculates a value to be stored in the current memory based on the current state and the input value of the memory cell. If it is 1, the previous hidden value is reflected as it is.

The update gate z determines whether to keep the old memory value as it is or to update it with a new value. If it is 1, the old memory is erased and a new value is used, and if 0, the old memory is used as it is.

In this case, the value can be expressed not only as 0 and 1, but also as a decimal number between 0 and 1 (eg 0.1, 0.54, 0.672, etc.).

1 is a block diagram of a personal diet recommendation service system according to an embodiment of the present invention.

As shown in FIG. 1 , the personal meal recommendation service system according to an embodiment of the present invention is configured to include a meal recommendation server 100 , a user terminal 200 , and a shopping mall terminal 300 .

The meal recommendation server 100 includes a meal recommendation learning unit 110, a display unit 120, and a meal information DB 130, and the meal recommendation learning unit 110 includes a time variable learning module 114 and a reflection learning module. 112 , and a recommendation module 116 .

As shown in FIG. 4, the diet recommendation learning unit 110 uses the meal the user ate immediately before in chronological order as an input layer, and uses the user's taste (eating) or pattern information as a hidden layer by deep learning in the GRU method. Then, it recommends (predicts) the diet.

At this time, the learning method can be learned by using menus such as kimchi stew and rolled egg rolls as one type of input layer, and can be learned separately for each ingredient rather than a menu, and meal times such as breakfast, lunch, and dinner can be divided. It can be learned, or it can be learned by including the user's taste (eating) or health information.

The time variable learning module 114 corresponds to the reset gate (r) of the GRU module shown in FIG. 2, and how the past (previous) pattern (diet) affects the future pattern (recommended diet) by deep learning determine whether

The reflection learning module 112 corresponds to the update gate z of the GRU module shown in FIG. 2 , and determines how much of the newly input ingredients and ingredients of the previous diet by deep learning are reflected in the future recommended diet.

The recommendation module 116 recommends a diet based on the reset variable of the time variable learning module 114 and the update variable of the reflection learning module 123 .

At this time, it is also possible to learn including the user's taste (eating) or health information and reflect this in the recommendation of the diet.

In addition, the recommendation module 116 may recommend various diets by setting priorities, and may find (match) a standard diet close to the recommended diet from the diet information DB 130 and provide it.

The diet information DB 130 stores various types of information necessary for the diet recommendation server 100 to recommend a diet, for example, user information (eating, health, pattern, etc.), diet information, standard diet information, food Includes recipes and ingredient information.

The display unit 120 displays the diet or matching standard diet recommended by the recommendation module 116, and when one of them is selected, the information (shopping mall name) of the food recipe of the corresponding diet or the ingredients of the corresponding diet is displayed. , location, phone number, price, etc.) can be displayed.

After accessing the meal recommendation server 100 and entering personal information (ID, password, diet, health, age, male and female, location, etc.) to join as a member, the personal diet provided by the meal recommendation server 100 The user terminal 200 may be further provided to use the recommendation service.

In this case, the meal recommendation server 100 may output and provide the contents displayed on the display unit 120 on the screen of the user terminal 200 .

A shopping mall terminal 300 may be further provided to connect to the meal recommendation server 100 and input food material information of nearby shops to help the meal recommendation server 100 recommend a meal while selling ingredients.

In addition, information on ingredients stored in the user's refrigerator 310 may be input to the food recommendation server 100 .

3 is a flowchart of a personal diet recommendation service method according to an embodiment of the present invention.

The time variable learning module 114 of the meal recommendation server 100 determines how much the past (previous) pattern (diet) affects the future pattern (recommended diet) by deep learning - past (previous) pattern (diet) ) to be applied to the future pattern (recommended diet) - is determined (S32).

The next meal recommendation server 100 receives new food material information from the outside (S33), and determines how much to reflect the newly input ingredients and the ingredients of the previous diet by deep learning in the reflection learning module 112 in the future recommended diet. (S34).

Next, the recommendation module 116 recommends a plurality of meals based on the reset variable of the time variable learning module 114 and the update variable of the reflection learning module 112 (S36).

In this case, the recommendation module 116 may find and match a standard diet close to the recommended diet from the diet information DB 130 .

Next, the meal recommendation server 100 outputs and provides a recommended meal or a matching standard meal on the screen of the user terminal 200 (S38).

On the other hand, before the step S32, the user terminal 300 accesses the meal recommendation server 100 and provides personal information, thereby affecting deep learning or meal recommendation of the meal recommendation server 100.

5 is a flowchart illustrating a learning method by deep learning of the present invention.

First, the training data required for deep learning among numerous past data is extracted (S51), and deep learning is performed with this training data (S52) to generate a new diet score model (S53).

Then, by simulating the past diet (S54) and comparing the performance of the existing model with the performance of the new diet score model (S55), if the performance of the new diet score model is better (S57), it is reflected as a recommendation model for the subsequent diet (S58) ), if the performance of the new diet score model is not better, new training data is extracted (S51).

6 is an exemplary view of a method for implementing a diet recommendation according to an embodiment of the present invention, in deep learning of the GRU model as shown in FIG. Kimchi stew, bean sprouts, and stewed tofu are recommended in the following diet.

In addition, as time passes, kimchi stew, bean sprouts, and braised tofu become the previous diet for the next meal, and bean sprouts, braised tofu, and cabbage kimchi are recommended in the subsequent diet by deep learning of the GRU model as shown in FIG. 3 .

Although the technical idea of the present invention has been described in conjunction with the accompanying drawings, the preferred embodiment of the present invention is exemplarily described and does not limit the present invention.

In addition, it is a clear fact that anyone with ordinary skill in the art can make various modifications and imitations within the scope of the technical idea of the present invention.

100: Meal recommendation server 110: Meal recommendation study unit
112: reflection learning module 114: time variable learning module
116: recommended module 120: display unit
130: diet information DB 200: user terminal
300: shopping mall terminal 310: refrigerator ingredient information

Claims (7)

A diet recommendation learning unit that uses the meal the user ate immediately before as the input layer and the user's eating habits or pattern information as the hidden layer, deep learning the previous data in the GRU (Gated Recurrent Unit) method, and recommends the next meal as the output layer. provided,
The GRU-type model introduces a reset gate and an update gate at each node of the hidden layer,
The diet recommendation learning unit includes a time variable learning module corresponding to the reset gate that determines how much the past diet affects the future recommended diet by deep learning, and the ingredients newly input by deep learning and ingredients of the previous diet into the future. A diet comprising: a reflection learning module corresponding to the update gate that determines how much to reflect in the recommended diet of Referral server. According to claim 1,
The food recommendation learning unit learns by using a food menu as a type of input layer, learns by dividing it by ingredients rather than a menu, learns by dividing breakfast, lunch, and dinner time, or learns the user's eating habits or health information Meal recommendation server, characterized in that learning, including. According to claim 1,
A menu information DB for storing user information, meal information, standard meal information, food recipe and ingredient information required for the meal recommendation server to recommend a meal, and a matching standard meal matching the meal recommended or recommended by the recommendation module A meal recommendation server, characterized in that the display unit for displaying is further provided. A meal recommendation server according to any one of claims 1 to 3;
a user terminal accessing the meal recommendation server, entering personal information to join as a member, and using a personal meal recommendation service provided by the meal recommendation server; and
a shopping mall terminal that connects to the meal recommendation server and supplies food material information of a nearby shopping mall to provide assistance when recommending a meal by the deep learning of the meal recommendation server;
A personal diet recommendation service system that includes. 5. The method of claim 4,
The meal recommendation server displays a recommended meal or a matching standard meal on the screen of the user terminal, and when one of them is selected, the food recipe of the corresponding meal or information of a shopping mall that sells the ingredients of the corresponding meal is displayed. A personal diet recommendation service system. (a) determining how much the past diet affects the future recommended diet by deep learning in the time variable learning module;
(b) receiving new ingredient information from the outside and determining how much to reflect the ingredients of the previous diet and newly input ingredients by deep learning in the reflection learning module in the future recommended diet; and
(c) recommending a diet based on the values determined in steps (a) and (b) in the recommendation module;
A personal diet recommendation service method that includes. 7. The method of claim 6,
A personal diet recommendation service method, characterized in that before the step (a), personal information such as personal diet or health information is provided.

Images