JP7077050B2 - Recommendation device, recommendation method and recommendation program - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Law COTOHA / AI Utilization Contest Grand Park Plaza Building 3F 301 Meeting Room Dates November 6, 2017, November 13, 2017 A Union Headquarters Date October 23, 2017 B Co., Ltd. Headquarters Date December 19, 2017 C Co., Ltd. Headquarters Date December 25, 2017

The present invention relates to a recommendation device, a recommendation method, and a recommendation program.

Conventionally, when a user is looking for a product such as wine, a technique for recommending the product according to the user's preference or the like has been known. As such a technique, for example, a technique for recommending a product according to a keyword specified by a user such as producer information, price, and product type is known. Also, for example, the user actually tastes some wines, and the user evaluates the tasted wines to adjust the criteria for the user's taste and recommend the wine according to the tendency toward the user's taste. There is a technology to do.

Japanese Unexamined Patent Publication No. 2014-140118

However, the conventional method has a problem that when a user requests a product recommendation with an ambiguous expression, the recommendation cannot be made. For example, there are individual differences in the taste and expression method for alcohol, and in keyword matching and the like, when a user requests a product recommendation with an ambiguous expression, the recommendation cannot be made. In the above-mentioned wine recommendation technique, it is premised that the user tastes the wine and adjusts the standard for the user's taste. Therefore, if the standard for the user's taste is not adjusted in advance, the recommendation is made. I couldn't do it.

In order to solve the above-mentioned problems and achieve the object, the recommendation device of the present invention analyzes a reception unit that accepts input of language data for a user to specify a product and a reception unit that accepts language data received by the reception unit. Then, a predetermined keyword is extracted from the language data, and a product recommended to the user from among the products stored in the storage unit that stores information about the product based on the keyword or a keyword having a meaning similar to the keyword. It is characterized by having a selection unit for selecting a product and an output unit for outputting information on the product selected by the selection unit.

Further, the recommendation method of the present invention is a recommendation method executed by a recommendation device, and analyzes a reception process for receiving input of language data for a user to specify a product and a language data received by the reception process. Then, a predetermined keyword is extracted from the language data, and a product recommended to the user from among the products stored in the storage unit that stores information about the product based on the keyword or a keyword having a similar meaning to the keyword. It is characterized by including a selection step of selecting a product and an output step of outputting information on the product selected by the selection step.

Further, the recommendation program of the present invention analyzes a reception step that accepts input of language data for a user to specify a product and the language data received by the reception step, and extracts a predetermined keyword from the language data. , A selection step of selecting a product recommended to the user from the products stored in the storage unit that stores information about the product based on the keyword or a keyword having a similar meaning to the keyword, and a selection step. It is characterized by having a computer execute an output step for outputting information on the product.

According to the present invention, even when a user requests a product recommendation with an ambiguous expression, it is possible to make an accurate recommendation.

FIG. 1 is a block diagram showing a configuration example of a recommendation device according to the first embodiment. FIG. 2 is a diagram showing an example of data stored in the sake information storage unit. FIG. 3 is a diagram showing an example of data stored in the manufacturer information storage unit. FIG. 4 is a diagram illustrating an example of outputting a question prompting the user for input. FIG. 5 is a diagram showing an example of output processing of recommended sake. FIG. 6 is a diagram illustrating a series of flow of recommendation processing by the recommendation device according to the first embodiment. FIG. 7 is a flowchart showing an example of the flow of the recommendation process in the recommendation device according to the first embodiment. FIG. 8 is a diagram showing a computer that executes a recommendation program.

Hereinafter, embodiments of the recommendation device, the recommendation method, and the recommendation program according to the present application will be described in detail with reference to the drawings. It should be noted that this embodiment does not limit the recommendation device, the recommendation method, and the recommendation program according to the present application.

[First Embodiment]
In the following embodiment, the configuration of the recommendation device 10 according to the first embodiment and the flow of processing of the recommendation device 10 will be described in order, and finally, the effect of the first embodiment will be described.

[Recommendation device configuration]
FIG. 1 is a block diagram showing a configuration example of a recommendation device according to the first embodiment. The configuration of the recommendation device 10 will be described with reference to FIG. As shown in FIG. 1, the recommendation device 10 is connected to the user terminal 20 via the network 30.

Here, the user terminal 20 is, for example, an information processing device such as a desktop PC, a tablet PC, a notebook PC, a mobile phone, a smartphone, a smart speaker, and a PDA (Personal Digital Assistant). The configuration shown in FIG. 1 is only an example, and the specific configuration and the number of each device are not particularly limited.

Further, as shown in FIG. 1, this recommendation device has a communication processing unit 11, a control unit 12, and a storage unit 13. The processing of each part of the recommendation device 10 will be described below.

The communication processing unit 11 controls communication related to various types of information. For example, the communication processing unit 11 receives language data for designating a product desired by the user from the user terminal 20. In addition, the communication processing unit 11 outputs information on the product recommended to the user. In the following, sake will be described as an example of the recommended product, but the product is not limited to this, and any product may be used, but the expression of the desired product differs depending on the user. It is assumed that the recommendation device 10 recommends products such as foods and drinks that are easy and easily ambiguous.

The storage unit 13 stores data and programs necessary for various processes by the control unit 12, and particularly closely related to the present invention, the storage unit 13 includes a sake information storage unit 13a and a manufacturer information storage unit 13b. For example, the storage unit 13 is a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk.

The sake information storage unit 13a stores information about sake. For example, as illustrated in FIG. 2, the sake information storage unit 13a associates the information related to sake with the "sake ID" that uniquely identifies sake, and associates it with the "brand" of sake, the "manufacturer" of sake, and sake. "Type" of sake, "price" of sake, "taste" of sake, "target user" indicating the attributes of users who like sake, "use" of sake, "label" of sake, "shape of bottle" of sake, It stores "additional information" which is additional information, "user evaluation" which shows user's evaluation, and "story" which shows the characteristics and background of sake. The information stored in the sake information storage unit 13a is updatable information. For example, when the sake with the sake ID "2" newly collaborates with △△, the information “collaborate with △△” can be added to the additional information. The information of each item is just an example and may be stored as a numerical value or a vector. For example, for the taste item, the degree of dryness or sweetness may be memorized, such as "dry: 3" or "sweet: 2".

The manufacturer information storage unit 13b stores information about the manufacturer of sake. For example, as illustrated in FIG. 3, the manufacturer information storage unit 13b has information about the manufacturer, such as "manufacturer" indicating the name of the manufacturer, "town with a brewery" indicating the name of the town where the sake brewery is located, and the manufacturer. The "history" that shows the number of years since the company was founded, the "Mori" that shows the name of Mori, and the "story" that shows the characteristics and background of the manufacturer's sake are stored in association with each other. The information stored in the sake information storage unit 13a and the manufacturer information storage unit 13b exemplified in FIGS. 2 and 3 is merely an example, and information other than the illustrated information may be stored, for example. The raw materials, manufacturing method, history and story of each sake, knowledge of the brewery, etc. may be stored as data.

The control unit 12 has an internal memory for storing a program that defines various processing procedures and required data, and executes various processing by these. However, the control unit 12 is particularly closely related to the present invention. It has a reception unit 12a, a selection unit 12b, an output unit 12c, and a feedback unit 12d. Here, the control unit 12 is an electronic circuit such as a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphical Processing Unit), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or the like. It is an integrated circuit.

The reception unit 12a accepts input of language data for the user to specify a product. For example, the reception unit 12a accepts input of language data for the user to specify sake. To explain with a specific example, for example, the reception unit 12a uses the user terminal 20 as language data such as "Dry Junmai Daiginjo", "Sake of the brand ○ Yama", and "Sake for beginners is good". , Accepts input of linguistic data such as "photogenic labels". The language data input here may be text data or voice data. When the input language data is voice data, for example, the reception unit 12a shall convert the voice data into text data by an existing method. Further, the language data to be input is not limited to Japanese, and may be a foreign language. If the input language data is a foreign language, for example, the reception unit 12a may translate the foreign language into Japanese using an existing translation tool or the like, or use the foreign language data model. The foreign language linguistic data input may be analyzed and converted into Japanese linguistic data.

Further, the reception unit 12a may output a question for causing the user to input predetermined language data, and may accept input of language data for the user to specify a product as an answer to the question. For example, the reception unit 12a narrows down the conditions for sake desired by the user by outputting a question to the user terminal 20 according to a preset format and repeating the process of receiving the answer to the question.

Here, an example of outputting a question prompting the user for input will be described with reference to FIG. FIG. 4 is a diagram illustrating an example of outputting a question prompting the user for input. As illustrated in FIG. 4, for example, the recommendation device 10 outputs a question "is it for a gift? Is it for home?" To the user terminal 20. Then, on the user terminal 20, the question "is it for a gift? Is it for home?" Is displayed as a text, or the voice "is it for a gift? Is it for home?" Is output.

On the other hand, when the user inputs the answer "for present" to the user terminal 20, the user terminal 20 sends the answer "for present" to the recommendation device 10. Note that, for example, the user terminal 20 may display a plurality of options for the question together with the question, and when the user taps any of the options, the recommendation device 10 uses the tapped option as the answer. You may send it to. Further, in the example of FIG. 4, the recommendation device 10 outputs the question "Who is the present?" To the user terminal 20.

Then, on the user terminal 20, the question "Who is the present?" Is output. On the other hand, when the user inputs the answer "I am a father" to the user terminal 20, the user terminal 20 sends the answer "I am a father" to the recommendation device 10.

Thereby, the recommendation device 10 can specify that the sake desired by the user is the sake for a gift and the recipient of the gift is the father. In the recommendation device 10, other expressions representing the father (for example, father, dad, etc.) are also set so that they can be recognized as synonymous with the father by using dictionary data and a language model. It is assumed that it has been done. Also, for example, although the user has specified only "for gifts", "father", and "sweet", there are many sakes that fall under the categories of "gift", "father", and "sweet" sake, so we recommend sake. In the case where the sake cannot be narrowed down, the recommendation device 10 can further output a question prompting the user to input, so that the condition of sake desired by the user can be narrowed down. Here, for questions that prompt the user to input in order to narrow down the conditions of sake desired by the user, the questions may be output sequentially according to a question pattern prepared in advance, or a model trained by AI or the like may be used. You may use it to output a question. Explaining the case where questions are sequentially output according to a question pattern prepared in advance by giving a specific example, for example, when the user is a present to the father, the recommendation device 10 is that the recipient of the present is the father. In some cases, as a pre-prepared question, the question "Which do you prefer, cold sake or hot sake?" Is further output. In addition, when the user specifies sweetness, the recommendation device 10 has various elements that can be narrowed down, such as asking how to drink, asking if you like sake, or asking if you are a beginner, as questions prepared in advance. It is possible to add questions.

The selection unit 12b analyzes the language data received by the reception unit 12a, extracts a predetermined keyword from the language data, and stores information about the sake based on the keyword or a keyword having a similar meaning to the keyword. A product recommended to the user is selected from the products stored in the information storage unit 13a. In addition, AI (Artificial Intelligence) may be made to execute the process of analyzing the language data.

For example, the selection unit 12b uses the learned model for analyzing the language data by inputting the language data received by the reception unit 12a, and recommends to the user from the sake stored in the sake information storage unit 13a. You may choose the sake you want to drink. When AI executes the process of analyzing language data, for example, AI performs morpheme analysis or the like on the language data input by the user to extract nouns or adjectives, and the extracted nouns or adjectives or extracts. Using a word similar to a noun or an adjective as a search key, the sake information storage unit 13a is referred to, and the sake corresponding to the same or similar data as the search key is selected. Note that AI can analyze synonymous words such as "sweet", "sweet", and "sweet" as words having the same or similar meanings by presetting or learning.

For example, when the reception unit 12a receives the linguistic data "Dry Junmai Daiginjo", the selection unit 12b refers to the sake information storage unit 13a, and the type is "Junmai Daiginjo". In addition, select sake that has a "sweet" taste. Further, for example, when the reception unit 12a receives the linguistic data "A mountain of sake brewing", the selection unit 12b refers to the sake information storage unit 13a, and the brand is "mountain ○". , Select sake whose manufacturer is "A Sake Brewery".

As described above, when the user likes sake and has a clear preference or is designated including a proper noun, the selection unit 12b searches for the sake information storage unit 13a and finds the corresponding unit. Select sake to do. If there are a plurality of applicable sakes, the selection unit 12b may select a plurality of the sakes, or may select one of them according to a preset priority or the like, or the user. You may try to narrow down the conditions by asking additional questions to.

Further, for example, when the reception unit 12a receives the linguistic data that "soft sake is good" and refers to the sake information storage unit 13a, the selection unit 12b corresponds to the ambiguous expression "fluffy". Does not exist in the sake information storage unit 13a. In such a case, the selection unit 12b analyzes the meaning of "fluffy" using a trained model for analyzing linguistic data, and for example, "fluffy" is a combination of "sweet" and "light". It is judged that it is a language to be expressed. Then, the selection unit 12b selects sake having a "sweet" taste and a "light" taste from the sake information storage unit 13a.

Further, for example, the selection unit 12b corresponds to the "photogenic label" which is an ambiguous expression when the reception unit 12a receives the linguistic data "photogenic label" and refers to the sake information storage unit 13a. Sake does not exist in the sake information storage unit 13a. In such a case, the selection unit 12b uses a trained model for analyzing linguistic data, and among the "photogenic labels", a similar expression having a similar meaning to the expression "photogenic". For example, a representative language is extracted from expressions such as "gorgeous", "beautiful", and "conspicuous", and the representative language is used as a search key from the "label" item in the sake information storage unit 13a. Select sake that matches or is similar to the representative search key. When the reception unit 12a receives the linguistic data of "photogenic sake", it analyzes the linguistic data of "photogenic sake" and relates to the appearance of the "label" and "bottle shape" of the sake. From the information, sake having a characteristic appearance may be selected from the sake information storage unit 13a.

Further, the selection unit 12b may select the related sake related to the sake information output by the output unit 12c from the sake information storage unit 13a. For example, the selection unit 12b identifies the most recently output sake from the history of the sake output by the output unit 12c, refers to the sake information storage unit 13a and the manufacturer information storage unit 13b, and refers to the same manufacturer as the most recently output sake. You may choose sake that has a sake brewery in the same town. Further, in the selection unit 12b, if the manufacturer of the most recently output sake is, for example, if the history is "100 years or older" or more, the history of the sake brewery is referred to as a series of places where the history of the sake brewery is located. You may choose an old sake that is "100 years old" or older.

The output unit 12c outputs information on sake selected by the selection unit 12b. For example, the output unit 12c outputs information such as an image of sake, a brand, a price, and a manufacturer as information on sake selected by the selection unit 12b. Here, the output processing of the recommended sake will be described with reference to the example of FIG. FIG. 5 is a diagram showing an example of output processing of recommended sake. As illustrated in FIG. 5, the recommendation device 10 receives the linguistic data "Tell me a fluffy sake" input by the user from the user terminal 20.

Then, the recommendation device 10 uses a trained model for analyzing linguistic data, and uses similar expressions that have a similar meaning to the expression "fluffy", for example, "sweet" or "light". By extracting a representative language, using the representative language as a search key, and selecting a sake that matches or is similar to the representative search key from the "taste" items in the sake information storage unit 13a, the taste can be changed. Select "sweet" and "light" sake "Daiginjo Oyama" from the sake information storage unit 13a. For similar languages, preset dictionary data may be used, or languages with short vector distances may be expressed as similar expressions by performing vector analysis. Subsequently, the recommendation device 10 outputs the linguistic data "How about Daiginjo Oyama?".

On the other hand, the user inputs the language data "it is better to be dry" to the user terminal 20, and the recommendation device 10 accepts the input of the language data. In this case, the recommendation device 10 selects the sake “Daiginjo Δ” having a “dry” taste from the sake information storage unit 13a. Subsequently, the recommendation device 10 outputs the linguistic data "How about Daiginjo △?".

Further, the output unit 12c outputs information on the related sake selected by the selection unit 12b. For example, if the manufacturer of the most recently output sake is, for example, an old sake with a history of "100 years" or more, the output unit 12c can be used as a series where the history of the sake brewery is "100 years" or more. You may output continuously. Further, the output unit 12c may output the sake selected by the selection unit 12b and output a story about the sake or a story about the manufacturer of the sake. For example, the output unit 12c may output the story of the sake manufacturer, etc., in addition to the information such as the image, brand, price, and manufacturer of sake.

When the feedback unit 12d receives language data as user feedback for the sake information output by the output unit 12c, the feedback unit 12d receives information about sake stored in the sake information storage unit 13a according to the language data. Update. For example, when the feedback unit 12d inputs the impression that the user is "easy to drink" with respect to the sake "○ mountain" output by the output unit 12c, the feedback unit 12d is "easy to drink" as a user evaluation for the sake "○ mountain". Is added to the sake information storage unit 13a.

Further, the feedback unit 12d may improve the recommendation accuracy by learning the feedback for the recommended sake information as learning data by AI. As described above, in the recommendation device 10, it is possible to improve the accuracy of the recommendation by receiving non-numerical feedback from the user by using the interactive interface.

Here, with reference to FIG. 6, a series of flow of the recommendation process by the recommendation device 10 will be described. FIG. 6 is a diagram illustrating a series of flow of recommendation processing by the recommendation device according to the first embodiment. As illustrated in FIG. 6, the recommendation device 10 receives input of language data for designating sake from the user terminal 20 (see (1) in FIG. 6). Then, the recommendation device 10 analyzes the language data by AI (see (2) in FIG. 6).

Subsequently, the recommendation device 10 selects a product recommended to the user from the products stored in the sake information storage unit 13a that stores information about sake based on the analysis result of the language data ((3) in FIG. 6 (3). )reference). Then, the recommendation device 10 outputs the information of the selected sake as the recommended sake (see (4) in FIG. 6). In this way, the recommendation device 10 can absorb fluctuations of various expressions included in the request from the user by AI and make a highly accurate recommendation.

In addition, the recommendation device 10 uses an interactive interface that can interact with AI so that even people who are not familiar with sake can easily recommend sake. For example, an example of a usage scene when a user wants to give sake as a gift will be described. For example, when the user tells that he wants to give sake as a gift, AI will give the gift to whom, what is the preference of the person to give the gift, etc. according to the question pattern preset as the pattern when the user gives sake as a gift. Ask the question.

Specifically, who the AI will give to, what kind of gift it is (for example, Father's Day, wedding gifts, etc.), what kind of gift is good (for example, photogenic, dry, good for beginners) Etc.) Ask a question. Then, the recommendation device 10 searches the sake information storage unit 13a using the result confirmed by AI as a keyword, and outputs one or a plurality of sake candidates to the user terminal 20. When a user purchases a product of interest from the output sake candidates, the product may be made available in cooperation with an EC (Electronic Commerce) site.

Further, for example, an example of a usage scene when the recommendation device 10 is used at a sake event will be described. For example, if the user tells AI the image and atmosphere he wants to drink (for example, clean, dry, for first-time drinkers, fruity, etc.), AI will select one sake that matches the image from the sake at the event venue. Or output multiple. Here, along with the information on sake, additional information such as the story of the sake and the location of the event venue may be presented.

Then, after the user drinks alcohol, AI asks for feedback, asks the user to input impressions such as "good" and "I like a little dry", and the input information is input to the sake information storage unit 13a. Store in the "User Evaluation" item. Then, if the user wishes to recommend the next sake, he / she may output the next recommended sake from the history information of the sake he / she has drank so far, or ask the user to input the request for the next sake. You may output the sake that you recommend next. Further, in the recommendation device 10, for example, when the same evaluation continues for a certain number of times, the information of the sake information storage unit 13a may be updated. For example, if a certain number of negative evaluations such as "a little dry is better" or "this is sweet" are counted as evaluations for sake recommended to users as dry sake, the target sake You may try to reduce the degree of dryness. It should be noted that the AI may determine how to reflect the feedback from the user in the information of the sake information storage unit 13a by learning.

In this way, with the recommendation device 10, even a person who is unfamiliar with sake can select the desired sake, such as when selecting sake for a gift, although he / she does not drink it by himself / herself. Even when a recommendation is requested, it is possible to make an accurate recommendation.

[Processing procedure for recommendation device]
Next, an example of the processing procedure by the recommendation device 10 according to the first embodiment will be described with reference to FIG. 7. FIG. 7 is a flowchart showing an example of the flow of the recommendation process in the recommendation device according to the first embodiment.

As illustrated in FIG. 7, when the reception unit 12a of the recommendation device 10 accepts the input of language data for designating sake from the user terminal 20 (step S101 affirmative), the selection unit 12b is accepted by the reception unit 12a. The linguistic data is analyzed (step S102), and based on the analysis result, a product recommended to the user is selected from the products stored in the sake information storage unit 13a that stores information about sake (step S103).

Then, the output unit 12c outputs the information of the sake selected by the selection unit 12b to the user terminal 20 (step S104). After that, the feedback unit 12d determines whether or not the user's feedback on the sake information output by the output unit 12c has been received from the user terminal 20 (step S105). As a result, if the feedback unit 12d does not accept the user's feedback (denial in step S105), the feedback unit 12d ends the process as it is. On the other hand, when the feedback unit 12d receives the user's feedback (step S105 affirmative), the feedback unit 12d stores the feedback in the sake information storage unit 13a (step S106).

(Effect of the first embodiment)
The recommendation device 10 according to the first embodiment accepts input of linguistic data for designating sake by a user, analyzes the received linguistic data, extracts a predetermined keyword from the linguistic data, and uses the keyword or the keyword. Even if the user requests a sake recommendation with an ambiguous expression based on a keyword having a similar meaning to the above, it is stored in the sake information storage unit 13a that stores information regarding the ability to make an accurate recommendation. Select the sake to recommend to the user from the sake, and output the information of the selected sake. Therefore, in the recommendation device 10, even when the user requests the recommendation of sake with an ambiguous expression, it is possible to make an accurate recommendation.

Further, in the recommendation device 10, it is possible to absorb fluctuations of various expressions included in the request from the user by AI and make a highly accurate recommendation. Further, the recommendation device 10 can easily recommend sake to a person who is not familiar with sake by using an interactive interface capable of interacting with AI.

(System configuration, etc.)
Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of them may be functionally or physically distributed / physically distributed in any unit according to various loads and usage conditions. Can be integrated and configured. Further, each processing function performed by each device may be realized by a CPU and a program analyzed and executed by the CPU, or may be realized as hardware by wired logic.

Further, among the processes described in the present embodiment, all or part of the processes described as being automatically performed can be manually performed, or the processes described as being manually performed. It is also possible to automatically perform all or part of the above by a known method. In addition, the processing procedure, control procedure, specific name, and information including various data and parameters shown in the above document and drawings can be arbitrarily changed unless otherwise specified.

(program)
It is also possible to create a program in which the processing executed by the recommendation device 10 described in the above embodiment is described in a language that can be executed by a computer. For example, it is also possible to create a recommendation program in which the processing executed by the recommendation device 10 according to the embodiment is described in a language that can be executed by a computer. In this case, the same effect as that of the above embodiment can be obtained by executing the recommendation program by the computer. Further, the same processing as that of the above embodiment may be realized by recording the recommendation program on a computer-readable recording medium and causing the computer to read and execute the recommendation program recorded on the recording medium.

FIG. 8 is a diagram showing a computer that executes a recommendation program. As illustrated in FIG. 8, the computer 1000 has, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. However, each of these parts is connected by a bus 1080.

The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM 1012, as illustrated in FIG. The ROM 1011 stores, for example, a boot program such as a BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1090, as illustrated in FIG. The disk drive interface 1040 is connected to the disk drive 1100 as illustrated in FIG. For example, a removable storage medium such as a magnetic disk or an optical disk is inserted into the disk drive 1100. The serial port interface 1050 is connected to, for example, a mouse 1110 and a keyboard 1120, as illustrated in FIG. The video adapter 1060 is connected, for example, to a display 1130, as illustrated in FIG.

Here, as illustrated in FIG. 8, the hard disk drive 1090 stores, for example, the OS 1091, the application program 1092, the program module 1093, and the program data 1094. That is, the above-mentioned recommendation program is stored in, for example, the hard disk drive 1090 as a program module in which a command executed by the computer 1000 is described.

Further, the various data described in the above embodiment are stored as program data in, for example, a memory 1010 or a hard disk drive 1090. Then, the CPU 1020 reads the program module 1093 and the program data 1094 stored in the memory 1010 and the hard disk drive 1090 into the RAM 1012 as needed, and executes various processing procedures.

The program module 1093 and program data 1094 related to the recommendation program are not limited to those stored in the hard disk drive 1090, and may be stored in, for example, a removable storage medium and read by the CPU 1020 via a disk drive or the like. good. Alternatively, the program module 1093 and the program data 1094 related to the recommendation program are stored in another computer connected via a network (LAN (Local Area Network), WAN (Wide Area Network), etc.), and are stored in another computer via the network interface 1070. It may be read by the CPU 1020.

The above-described embodiments and modifications thereof are included in the invention described in the claims and the equivalent scope thereof, as included in the technique disclosed in the present application.

10 Recommendation device 11 Communication processing unit 12 Control unit 12a Reception unit 12b Selection unit 12c Output unit 12d Feedback unit 13 Storage unit 13a Sake information storage unit 13b Manufacturer information storage unit 20 User terminal 30 Network

Claims (7)

A reception desk that accepts input of language data for users to specify sake,
Information on the manufacturer of sake, along with information on sake, is analyzed by analyzing the language data received by the reception unit, extracting a predetermined keyword from the language data, and based on the keyword or a keyword having a similar meaning to the keyword. As for the manufacturer that shows the name of the manufacturer, the town that has the brewery that shows the name of the town where the sake brewery is located, the history that shows the number of years since the manufacturer was founded, and the sake that shows the name of Mr. Mori. A selection unit that selects sake to be recommended to the user from among the sake stored in the storage unit that stores one or more of the stories related to.
It has an output unit that outputs information about the manufacturer corresponding to the sake, as well as information on the sake selected by the selection unit.
The selection unit identifies the most recently output sake from the information of the sake output by the output unit, refers to the storage unit, and has the same manufacturer, the same forest, or the same as the most recently output sake. A recommendation device that features the selection of sake that has a sake brewery in the town . The reception unit outputs a question for causing the user to input predetermined language data, and as an answer to the question, the reception unit accepts input of language data for designating sake by the user according to claim 1. The recommended device described. The selection unit receives the language data received by the reception unit as an input, and uses a trained model for analyzing the language data to recommend sake to the user from among the sake stored in the storage unit. The recommendation device according to claim 1, wherein the recommendation device is selected. The selection unit selects related sake related to the sake information output by the output unit from the storage unit.
The recommendation device according to claim 1, wherein the output unit outputs information on related sake selected by the selection unit. As user feedback on the sake information output by the output unit, when the language data is received, a feedback unit that updates the information on sake stored in the storage unit according to the language data is further provided. The recommendation device according to claim 1, wherein the recommendation device is provided. A recommendation method performed by a recommendation device,
The reception process that accepts the input of language data for the user to specify sake,
The language data received by the reception process is analyzed, a predetermined keyword is extracted from the language data, and based on the keyword or a keyword having a similar meaning to the keyword, information on sake and information on the manufacturer of the sake are obtained. As for the manufacturer that shows the name of the manufacturer, the town that has the brewery that shows the name of the town where the sake brewery is located, the history that shows the number of years since the manufacturer was founded, and the sake that shows the name of Mr. Mori. A selection process for selecting sake to be recommended to the user from among the sake stored in the storage unit that stores one or more of the stories related to the above.
Includes an output process that outputs information about the manufacturer corresponding to the sake, as well as information on the sake selected by the selection process.
The selection step identifies the most recently output sake from the information of the sake output by the output step, refers to the storage unit, and has the same manufacturer, the same forest, or the same as the most recently output sake. A recommendation method characterized by selecting sake that has a sake brewery in the town . A reception step that accepts input of language data for the user to specify sake,
The language data received by the reception step is analyzed, a predetermined keyword is extracted from the language data, and based on the keyword or a keyword having a similar meaning to the keyword, information on sake and information on the manufacturer of the sake are obtained. As for the manufacturer that shows the name of the manufacturer, the town that has the brewery that shows the name of the town where the sake brewery is located, the history that shows the number of years since the manufacturer was founded, and the sake that shows the name of Mr. Mori. A selection step to select the sake recommended to the user from the sake stored in the storage unit that stores one or more of the stories related to.
The computer is made to execute an output step of outputting information about the manufacturer corresponding to the sake together with the information of the sake selected by the selection step .
The selection step identifies the most recently output sake from the information of the sake output by the output step, refers to the storage unit, and has the same manufacturer, the same forest, or the same as the most recently output sake. A recommendation program that features the selection of sake that has a sake brewery in the town .

Images