KR20210110006A - Server and method for recommending product - Google Patents

Abstract

A product recommendation device comprises: a keyword collection unit for collecting a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of pieces of additional information from a plurality of websites; a vector generation unit for generating an item vector for each of the products on the basis of the first keywords and respective frequencies of the first keywords, and generating an additional vector for each of the pieces of additional information on the basis of the second keywords and respective frequencies of the second keywords; a request reception unit for receiving a product recommendation request including information on a first product and first additional information from a user terminal; and a product recommendation unit for recommending a plurality of products different from the first product to the user terminal on the basis of an item vector for the first product among item vectors for the respective products and an additional vector for the first additional information among additional vectors for the respective pieces of additional information. A personalized product can be recommended to a user terminal.

Description

SERVER AND METHOD FOR RECOMMENDING PRODUCT

The present invention relates to a server and method for recommending a product.

Recently, with the development of the IT industry, a system for recommending personalized services or products has been developed in various forms.

In general, the recommendation system is a personalized information filtering technology that helps customers easily find products that they are likely to purchase by creating a list of recommended products for customers.

Collaborative filtering used in the existing recommendation system is one of the data mining techniques, and in particular, a user data analysis technique. Most of the recommended products on Internet shopping mall sites are derived through collaborative filtering.

However, in the recommendation system using collaborative filtering, as the number of products and customers managed by the Internet shopping mall rapidly increases, the problem of sparsity of input data and the problem of system scalability are occurring.

In addition, in the case of individual operators who promote business through Internet shopping malls or apps, there are technical barriers to conducting business through a recommendation system. In order for non-specialists who do not have specialized data mining skills to provide personalized recommendation services using a recommendation system in various fields, a method to lighten the recommendation system development process is needed.

On the other hand, the existing personalized product recommendation uses personal information such as the user's age, gender, and current location to recommend products purchased by nearby users similar to or identical to the user's personal information to the user. In order to do this, the user's personal information is absolutely necessary. However, there is a problem in that it is difficult to freely use the user's personal information in order to protect the user's personal information.

Korean Patent Publication No. 2010-0132824 (published on December 20, 2010)

The present invention is to solve the problems of the prior art described above, and based on a vector for each product and additional information included in a product recommendation request received from a user terminal, to recommend a plurality of products different from the corresponding product to the user terminal. do. However, the technical problems to be achieved by the present embodiment are not limited to the technical problems described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, the product recommendation apparatus according to the first aspect of the present invention provides a plurality of first keywords and a plurality of additional information for each of a plurality of products from a plurality of websites. a keyword collecting unit collecting a second keyword; generating an item vector for each of the plurality of products based on the plurality of first keywords and the respective frequencies of the plurality of first keywords, and based on the respective frequencies of the plurality of second keywords and the plurality of second keywords a vector generator for generating an additional vector for each of the plurality of pieces of additional information; a request receiving unit for receiving a product recommendation request including information on a first product and first additional information from the user terminal; and a plurality of items different from the first product based on the additional vector for the first additional information among the item vector for the first product among the item vectors for each of the plurality of products and the additional vector for each of the plurality of additional information It may include a product recommendation unit for recommending a product of the user terminal.

A product recommendation method according to a second aspect of the present invention includes: collecting a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of additional information from a plurality of websites; generating an item vector for each of the plurality of products based on the plurality of first keywords and the respective frequencies of the plurality of first keywords, and based on the respective frequencies of the plurality of second keywords and the plurality of second keywords generating an additional vector for each of the plurality of pieces of additional information; receiving a product recommendation request including information on a first product and first additional information from a user terminal; and a plurality of items different from the first product based on the additional vector for the first additional information among the item vector for the first product among the item vectors for each of the plurality of products and the additional vector for each of the plurality of additional information It may include the step of recommending a product of the user terminal.

The above-described problem solving means are merely exemplary, and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, there may be additional embodiments described in the drawings and detailed description.

According to any one of the above-described problem solving means of the present invention, the present invention can recommend a personalized product to the user terminal based on the vector for each product and additional information included in the product recommendation request received from the user terminal. have.

1 is a block diagram of a product recommendation apparatus according to an embodiment of the present invention.
2A to 2B are diagrams illustrating tables storing keywords and frequencies for a plurality of products, respectively, according to an embodiment of the present invention.
3A to 3B are diagrams illustrating tables storing keywords and frequencies for a plurality of pieces of additional information, respectively, according to an embodiment of the present invention.
4 is a flowchart illustrating a method of recommending a product according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art can easily implement them. However, the present invention may be embodied in many different forms and is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is "connected" with another part, this includes not only the case of being "directly connected" but also the case of being "electrically connected" with another element interposed therebetween. . 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 this specification, a "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. In addition, one unit may be implemented using two or more hardware, and two or more units may be implemented by one hardware.

Some of the operations or functions described as being performed by the terminal or device in the present specification may be instead performed by a server connected to the terminal or device. Similarly, some of the operations or functions described as being performed by the server may also be performed in a terminal or device connected to the server.

Hereinafter, detailed contents for carrying out the present invention will be described with reference to the accompanying configuration diagram or process flow diagram.

1 is a block diagram of a product recommendation apparatus 10 according to an embodiment of the present invention.

Referring to FIG. 1 , the product recommendation apparatus 10 may include a keyword collection unit 100 , a vector generation unit 110 , a request receiving unit 120 , and a product recommendation unit 130 . However, the product recommendation device 10 shown in FIG. 1 is only one embodiment of the present invention, and various modifications are possible based on the components shown in FIG. 1 .

Hereinafter, FIG. 1 will be described with reference to FIGS. 2A to 3B.

The keyword collection unit 100 may collect a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of additional information from a plurality of websites. Here, the plurality of additional information is all information that affects product recommendation, for example, environmental information (eg, weather information, season information, etc.), country of origin information, ingredient information, related product information, and recommendation purpose information (for gifts). , for decoration, etc.) and the like.

The keyword collecting unit 100 calculates the frequency of occurrence of each of the plurality of first keywords for each of the plurality of products from the plurality of websites, and each of the plurality of second keywords for each of the plurality of additional information from the plurality of websites The frequency of this appearance can be calculated.

The keyword collection unit 100 may store a table in which a first keyword of each product and a frequency number of each first keyword are mapped for each product. Also, the keyword collection unit 100 may store a table in which the second keyword of each additional information and the frequency of each of the second keywords are mapped for each additional information.

For example, referring to FIGS. 2A to 2B , the keyword collection unit 100 may crawl keywords related to the first wine.

For example, the keyword collection unit 100 may include 'delicious' keywords, 'sweet' keywords, 'talk' related to the first wine from a plurality of websites corresponding to the search results for the first wine searched through the search engine. It is possible to collect a plurality of first keywords including the 'shoot' keyword and the 'fresh' keyword, and calculate the frequency of each of the plurality of first keywords appearing on the plurality of web sites. For example, the keyword collection unit 100 calculates the frequency of occurrence of the keyword 'delicious' related to the first wine in a plurality of web sites corresponding to the search results for the first wine, and 'in the information of the first wine' You can map the frequency of 'delicious' keywords and 'delicious' keywords and store them in a table.

For another example, the keyword collection unit 100 may collect keywords for the first wine from a review site of a preset product, social network services (SNS), and other communities.

In addition, the keyword collection unit 100 is 'delicious' keyword, 'bitter' keyword, 'fresh' keyword, 'plain' related to the second wine from a plurality of websites corresponding to the search result for the second wine. A plurality of first keywords including the keyword are collected, and the frequency of each of the plurality of first keywords is calculated in a plurality of websites (websites where the second product is searched for), and each first keyword is added to the information of the second wine. A keyword and the frequency of each first keyword may be mapped and stored in a table.

For another example, the keyword collection unit 100 may collect keywords for the second wine from a review site of a preset product, SNS, and other communities.

For example, referring to FIGS. 3A to 3B , the keyword collecting unit 100 provides weather information from a plurality of websites (eg, 'I want to drink the first wine in sunny weather' or 'I drink the first wine in cloudy weather, If the weather information extracted from 'I want', etc.) is extracted, a plurality of second keywords are collected for each of the weather information (eg, 'sunny', 'flow') from a plurality of websites, It is possible to calculate the frequency of each second keyword. For example, the keyword collection unit 100 may include emotional keywords (eg, feel good), travel keywords (eg, I want to go on a trip), walking keywords (eg, I feel good) in relation to weather information corresponding to 'sunny' from a plurality of web pages. For example, it is possible to collect a plurality of second keywords including, for example, I want to take a walk) and music keywords (eg, I want to listen to bright music). In addition, the keyword collection unit 100 may include an emotional keyword (eg, depressed), a home keyword (eg, want to be at home) and a music keyword (eg, I want to listen to soothing music) and the like) can be collected. Here, the plurality of websites may include SNS, a plurality of communities, and the like.

The vector generator 110 may generate an item vector for each of the plurality of products based on the plurality of first keywords for each of the plurality of products and the frequency of each of the plurality of first keywords.

For example, referring to FIG. 2A , the vector generating unit 110 is based on the frequency of each of the 'delicious' keyword, 'sweet' keyword, 'tickle' keyword, and 'fresh' keyword for the first wine. An item vector for the first wine may be created.

Referring to FIG. 2B , the vector generating unit 110 for the second wine based on the frequency of each of the keyword 'delicious', 'bitter', 'fresh' keyword, and 'plain' keyword for the second wine. You can create item vectors.

The vector generator 110 may generate an additional vector for each of the plurality of additional information based on the plurality of second keywords for each of the plurality of additional information and the frequency of each of the plurality of second keywords.

For example, referring to FIGS. 3A to 3B , the vector generating unit 110 generates a second based on the frequency of each of the emotion keyword, travel keyword, walking keyword, and music keyword for the first weather information (eg, 'sunny'). 1 An additional vector for weather information may be generated. Also, the vector generator 110 may generate an additional vector for the second weather information based on the frequency of each of the emotion keyword, the home keyword, and the music keyword for the second weather information (eg, 'cloudy').

The request receiving unit 120 may receive a product recommendation request including information on the first product and the first additional information from the user terminal. For example, the request receiving unit 120 may receive a product recommendation request including 'recommend fresh wine to drink in clear weather' from the user terminal.

When the first additional information included in the product recommendation request is collected in advance, the vector generator 110 extracts an item vector for the first product from among the item vectors for each of a plurality of products stored in the table, and stores the item vector stored in the table. An additional vector for the first additional information may be extracted from among the additional vectors for each of the plurality of additional information.

The product recommendation unit 130 is configured to obtain a second product based on the item vector for the first product extracted from the item vector for each of the plurality of products and the additional vector for the first additional information extracted from the additional vector for each of the plurality of additional information A plurality of products different from the first product may be recommended to the user terminal. For example, the product recommendation unit 130 may recommend a plurality of different wines to the user terminal based on the item vector for 'fresh wine' and the additional vector for 'clear weather' additional information.

The product recommendation unit 130 may recommend a plurality of different products to the user terminal by reflecting a preset weight to each of the item vector for the first product and the additional vector for the first additional information. Here, the preset weight is for which vector among the item vector for the first product and the additional vector for the first additional information is to be given a greater weight, and the user's feedback ratio according to the product recommendation to the user terminal or the new It may be changed based on the presence or absence of additional information. For example, the product recommendation unit 130 may reflect a first weight to the item vector for 'fresh wine' and to reflect the second weight to the additional vector for 'clear weather' additional information.

The product recommendation unit 130 may generate a new vector based on the item vector for the first product and the additional vector for the first additional information. Also, the product recommendation unit 130 may generate a new vector by using preset weights reflected in each of the item vector for the first product and the additional vector for the first additional information. Here, the product recommendation unit 130 may generate a new vector based on [Equation 1].

[Equation 1]

New vector = α*(item vector for first product) + (1-α)*(addition vector for first additional information)

Here, α is a preset weight, and 0<α<1.

The product recommendation unit 130 may extract an item vector similar to the new vector from among the item vectors for each of a plurality of products stored in the table, and recommend a product corresponding to the similar item vector to the user terminal as one of a plurality of other products. .

Here, the vector generator 110 may calculate the similarity between the new vector and the item vector of each of a plurality of other products having an item vector similar to the new vector through [Equation 2]. Here, the similarity may be a cosine similarity.

[Equation 2]

The product recommendation unit 130 may recommend a plurality of other products having an item vector similar to the new vector in order of similarity with the new vector to the user terminal. For example, the product recommendation unit 130 may sort the similarity with the new vector in a descending order and recommend the product to the user terminal in the order of the high similarity to the new vector.

When the additional vector for the first additional information included in the product recommendation request is not collected in advance, the keyword collection unit 100 collects a plurality of third keywords for the first additional information from a preset website in real time, , it is possible to calculate the frequency of appearance of each third keyword from a preset website.

In addition, the keyword collection unit 100 maps a plurality of third keywords for the first additional information and the frequency of each third keyword to the first product based on the information on the first product included in the product recommendation request to create a new You can additionally store them in a table. In the future, the keyword collection unit 100 may update the corresponding new table through the above-described crawling.

The vector generator 110 may generate an additional vector for the first additional information in real time based on the plurality of third keywords for the first additional information and the frequency of each of the plurality of third keywords.

The product recommendation unit 130 may recommend a plurality of products different from the first product to the user terminal based on the item vector for the first product and the additional vector for the first additional information generated in real time.

The product recommendation unit 130 may recommend a plurality of other products to the user terminal by reflecting a higher weight in the item vector for the first product than the additional vector for the first additional information generated in real time. This is because the additional vector for the first additional information is generated in real time in a short period of time, so the reliability is low.

When the product recommendation unit 130 receives user feedback information from a plurality of user terminals in relation to a plurality of products different from the first product recommended to the plurality of user terminals, the item vector based on the feedback information of the plurality of users and a preset weight to be given to the additional vector may be adjusted. For example, the product recommendation unit 130 calculates an average value for the feedback information of a plurality of users, and when the calculated average value is greater than or equal to a preset threshold value, the existing preset weight is not changed, and the calculated average value is When it is less than or equal to a set threshold, an existing preset weight may be changed.

On the other hand, those skilled in the art will recognize that each of the keyword collection unit 100, the vector generation unit 110, the request receiving unit 120, and the product recommendation unit 130 may be implemented separately, or that one or more of them may be integrated. will understand enough.

4 is a flowchart illustrating a method of recommending a product according to an embodiment of the present invention.

Referring to FIG. 4 , in step S401 , the product recommendation apparatus 10 may collect a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of additional information from a plurality of websites. .

In step S403 , the product recommendation apparatus 10 generates an item vector for each of the plurality of products based on the plurality of first keywords and the respective frequencies of the plurality of first keywords, and the plurality of second keywords and the plurality of second keywords An additional vector for each of the plurality of pieces of additional information may be generated based on each frequency.

In step S405 , the product recommendation apparatus 10 may receive a product recommendation request including information about the first product and the first additional information from the user terminal.

In step S407, the product recommendation apparatus 10 performs a first step based on the additional vector for the first additional information among the item vector for the first product among the item vectors for each of the plurality of products and the additional vector for the first additional information among the additional vectors for each of the plurality of additional information A plurality of products different from the product may be recommended to the user terminal.

In the above description, steps S401 to S407 may be further divided into additional steps or combined into fewer steps, according to an embodiment of the present invention. In addition, some steps may be omitted as necessary, and the order between steps may be changed.

An embodiment of the present invention may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include all computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The description of the present invention described above is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention. .

10: Product recommendation device
100: keyword collection unit
110: vector generator
120: request receiving unit
130: product recommendation unit

Claims (10)

In the product recommendation device,
a keyword collecting unit for collecting a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of additional information from a plurality of websites;
generating an item vector for each of the plurality of products based on the plurality of first keywords and the respective frequencies of the plurality of first keywords, and based on the respective frequencies of the plurality of second keywords and the plurality of second keywords a vector generator for generating an additional vector for each of the plurality of pieces of additional information;
a request receiving unit for receiving a product recommendation request including information on a first product and first additional information from the user terminal; and
A plurality of items different from the first product on the basis of an item vector for the first product among the item vectors for each of the plurality of products and an additional vector for the first additional information among the additional vectors for each of the plurality of additional information A product recommendation unit that recommends a product to the user terminal
Which will include, a product recommendation device.
The method of claim 1,
wherein the product recommendation unit recommends the other plurality of products by reflecting a preset weight to each of the item vector for the first product and the additional vector for the first additional information.
The method of claim 1,
The product recommendation unit generates a new vector based on the item vector for the first product and the additional vector for the first additional information, and extracts an item vector similar to the new vector from the item vectors for each of the plurality of products and recommending a product corresponding to the similar item vector as one of the other plurality of products.
4. The method of claim 3,
The product recommendation section
The product recommendation apparatus of claim 1, wherein the plurality of other products are arranged in the order of similarity with the new vector and recommended to the user terminal.
The method of claim 1,
If the additional vector for the first additional information has not been collected in advance, the keyword collection unit collects a plurality of third keywords for the first additional information from a preset website in real time,
wherein the vector generator generates an additional vector for the first additional information in real time based on the plurality of third keywords and the frequency of each of the plurality of third keywords.
6. The method of claim 5,
wherein the product recommendation unit reflects a higher weight to the item vector for the first product than to the additional vector for the first additional information.
5. The method of claim 4,
The similarity will be a cosine similarity, product recommendation device.
In the product recommendation method,
collecting a plurality of first keywords for each of a plurality of products and a plurality of second keywords for each of a plurality of additional information from a plurality of websites;
generating an item vector for each of the plurality of products based on the plurality of first keywords and the respective frequencies of the plurality of first keywords, and based on the respective frequencies of the plurality of second keywords and the plurality of second keywords generating an additional vector for each of the plurality of pieces of additional information;
receiving a product recommendation request including information on a first product and first additional information from a user terminal; and
A plurality of items different from the first product on the basis of an item vector for the first product among the item vectors for each of the plurality of products and an additional vector for the first additional information among the additional vectors for each of the plurality of additional information Recommending a product to the user terminal
A method for recommending a product, comprising:
9. The method of claim 8,
The step of recommending a plurality of products different from the first product to the user terminal
and recommending the other plurality of products by reflecting a preset weight to each of the item vector for the first product and the additional vector for the first additional information.
9. The method of claim 8,
collecting, in real time, a plurality of third keywords for the first additional information from a preset website when the additional vector for the first additional information has not been previously collected; and
generating an additional vector for the first additional information in real time based on the plurality of third keywords and the frequency of each of the plurality of third keywords
Which will further include, a product recommendation method.

Images