JP2009140032A - Clothing recommending system according to user's circumstances - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a clothes recommendation system allowing an assumption of a specific situation of a user, and allowing input by a natural language sentence not but a keyword so as to solve the problem, wherein multi-retrieval is obliged, until finding a target commodity, because the user has to always guess the category or the keyword related to the commodity, in keyword retrieval and category retrieval which is presently widely adopted on a shopping site over the Internet. <P>SOLUTION: In this clothes recommendation system 16, based on the situation of the user, the natural language sentence inputted from the user 15 is morpheme-analyzed in a mechanism 9, an index of an according word, and a range and an attribute of the index are extracted in a mechanism 10, and clothing according with the index, or the like, is extracted from a clothes information DB 2 in a mechanism 11. Finally, the recommended clothes are divided into categories in a mechanism 12 and are displayed on a screen. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

The present invention extracts words contained in a word dictionary from a user's natural language sentence input from a clothes database in which six indices are added to each garment, and sets indices that match the indices added to the above words. The present invention relates to a clothing recommendation system that recommends clothing to be held.

Currently, there are many shopping sites on the Internet. These shopping sites handle various categories (types) of products and are used by many users.
Many of such shopping sites employ a keyword search for searching for products using keywords related to products and a category search for searching for products by category.

However, the shopping site as described above requires a plurality of searches in order to guess the keyword of the product when the product to be purchased by the user is not clear and find the target product from a large-scale product database. A new mechanism for solving the above problems has been demanded.
There are also shopping sites that have introduced technologies such as emphasis filtering that recommend products to users who have purchased similar products from the purchase history of each user. Since the mechanism is not considered, the burden on the user when searching for the product has not been resolved.

Therefore, the present invention proposes the following method and system for the purpose of providing a clothing recommendation system that inputs a user's situation in a natural language sentence and recommends a product by the above input.
That is, in claim 1, the user inputs the situation in a natural language sentence to the input form. In the method of recommending clothes from an input sentence, the system consists of a word dictionary containing words indicating events and people, a clothes database, an input part, an output part, and an avatar part. Extract and recommend this.
In claim 2, the input by the user of claim 1 inputs a specific situation of participating in some event. Examples of input include “going to a meal with the boss” and “participating in the daughter's entrance ceremony”. The user's situation is specified from the above input, and clothing recommendation suitable for the situation is executed.
According to the third aspect of the present invention, the morphological analysis is executed in order for the system to grasp the word included in the input sentence and its part of speech.
Six indicators added to each garment of claim 4 are a formal degree, a fashion degree, a sporty degree, a gorgeous degree, an elegance degree, and an individuality degree, respectively. Each index is indicated by a 10-level numerical value. Each word in the word dictionary has a formal level, a range of the formal level, and an attribute indicating an event or a person. The input sentence and clothes are matched using the above.
Claims 5 and 6 are recommended when the index of the corresponding word extracted in claim 4, for example, the formal degree of meal is “4”, the range of formal degree is “4 or less”, and the attribute is “event”. The product will be clothing with a formal degree of "4 or less".
In claim 7, by selecting the clothes recommended by the present system, the combination of clothes can be freely confirmed in the avatar part.
In claim 8, the formal degree of the event may change depending on the partner who spends the event. For example, when comparing “eating with family” and “eating with boss”, “eating with boss” is a more formal event. In this way, it is possible not only to focus on events, but also to recommend clothes according to the person you spend.
According to the ninth aspect of the present invention, it is possible to confirm a combination of items already purchased by the user with clothes newly purchased in the avatar part.

According to the first and second aspects, a specific situation of a user who uses the system is assumed, and clothes having properties suitable for the inputted situation are searched from the clothes database, and the above is recommended. Unlike conventional product search systems, there is no need to search by guessing keywords or categories, and the user can search for clothes simply by inputting the situation.
According to the third aspect, the morphological analysis is performed in order to extract a word indicating an event or a partner to spend from an input sentence from the user. By dividing the input sentence into words, it is possible to easily extract an index of words used when recommending clothes.
According to the fourth, fifth, and sixth aspects, a word indicating a partner who spends an event or an event is searched from the word dictionary from the analysis result divided into words by morphological analysis, and an index added to the matching word is extracted. Furthermore, by searching for matching clothes using the above index in the clothes database, it is possible to recommend clothes that match the situation entered by the user, and it is possible to search for clothes even if the clothes that the user wants to purchase are not clearly determined System.
According to the seventh and ninth aspects, the user can freely coordinate and check the combination of the recommended clothes and the purchased clothes as well as recommending the clothes, and the user can select the combination of the clothes that the user likes before purchasing the clothes. It is a cheap system that can be confirmed.

Here, the best mode for carrying out the present invention will be described with reference to the drawings. FIG. 1 shows a basic configuration of a clothing recommendation system based on a user's situation according to the present invention.
The clothing recommendation system 16 according to the user's purpose includes an input unit 7, a mechanism 9 for performing morphological analysis, a mechanism 10 for searching for a word indicating an event or a partner to spend with an input sentence, a mechanism 11 for searching for matching clothes It comprises a mechanism 12 for classifying recommended clothes, an output unit, and an avatar unit 10.
As for the input part 7, a system receives an input sentence, when a user inputs a condition.
In the mechanism 9, the morphological analysis is performed on the input sentence received from the input unit by the morphological analysis module 13, and as a result of the analysis, a word and a part of speech of the word are extracted from the input sentence (processing result 3).
The mechanism 10 searches from a word dictionary that matches the word extracted from the mechanism 9. Here, the index added to the matched word, the range of the index, and the attribute are extracted (processing result 4).
The mechanism 11 searches for matching clothes from the clothes information DB 2 using the processing result 4 of the mechanism 10. The matched clothes are temporarily stored as the processing result 5 regardless of the category (jacket, shirt, trousers).
The mechanism 12 categorizes the processing result 5 of the mechanism 11. The recommended clothing after categorization is set as a processing result 6.
The output unit and the avatar unit 10 receive the processing result 5 of the mechanism 12 and display the above recommended clothing. The selected clothes are displayed larger than usual and can be coordinated with other assistants.
FIG. 2 is a diagram showing a flow until the user receives a clothing recommendation. User input input to the input unit is morphologically analyzed and classified into parts of speech and original forms of words. When there is a match between the above word and the word dictionary, the index of the word, the range of the index, and the attribute are extracted. Next, a matching clothing is searched from the clothing information DB for a series of extracted word data, and the matching clothing is displayed as a recommended product. Finally, the user confirms the combination in the avatar part. If you don't like it, you can change the index and search again.

It is a figure which shows the basic composition of the clothing recommendation system based on the condition of the user by this invention. It is a flowchart until a user receives clothes recommendation. It is a figure which shows the data structure of a word dictionary. It is a figure which shows the data structure of a clothing information database.

Explanation of symbols

1 User 2 Clothes information database (DB)
3, 4, 5, 6, processing result 7 input 8 output 9 mechanism for performing morphological analysis 10 mechanism for extracting index from words 11 mechanism for searching for clothes 12 mechanism for categorizing recommended clothes 13 morphological analysis module 14 using index Clothes search 15 user 16 clothes recommendation system

Claims (9)

A clothing recommendation system that recommends clothing suitable for a user's situation based on a situation input by the user from a database storing clothing information. The clothing recommendation system according to claim 1, wherein the user inputs a person who participates in an event to be attended in the future, assuming a specific situation in which the user participates in some event. 2. The clothing recommendation system according to claim 1, wherein the input sentence is subjected to morphological analysis and the analysis result is used for searching for clothing. 6. A clothing recommendation system using a word dictionary in which six indexes representing properties are added to each clothing registered in the clothing database according to claim 1 and further indexes are added to words indicating events and people. . A clothing recommendation system, wherein a word contained in the analysis result of claim 3 is searched from the word dictionary of claim 4 and an index added to the matched word is used for clothing search. 6. A clothing recommendation system, wherein clothing matching the index of the word matched in claim 5 is extracted from the clothing database of claim 4. The clothes recommendation system according to claim 6, wherein the extracted clothes are output on a screen, and the user can freely coordinate the clothes in the avatar part. 4. The clothing recommendation system according to claim 3, wherein the recommended clothing is changed by a person accompanying the user when participating in the event. A clothing recommendation system that can be freely coordinated with the clothing recommended by the system and the avatar by registering in advance the products held by the user.