JP2002269106A - Device for introducing book - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a book introducing device capable of efficiently providing books that can be interesting to a user on the basis of preference of the user by automatically deciding the user preference from the detailed description information of books read by the user and other interactive inputs. SOLUTION: This book introducing device 1 is provided with: a user information database 22 for storing and managing user personal information and preference information; a book information database 21 for storing and managing book information; an interactive processing controlling means 17 for receiving a retrieval request from the terminal device of the user, starting a necessary processing means and answering results to the user terminal; a book recommending means 14 for retrieving the book information database while referring to the preference information of the user and recommending and presenting a book coinciding with the user preference ; and a user profile updating means 15 for updating user preference information data from an interactive input of the user.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a book introduction device to which a technology for recommending information is applied.

[0002]

2. Description of the Related Art It is becoming common to purchase and sell products in virtual stores established on the Internet for general consumers and consumers. Some bookstores have been set up in virtual space. Such a bookstore is provided with a database, and when a user tries to find a target book, the target book can be designated by performing a database search. In addition, in order to help the user accidentally search for a book, for a selected book, the book purchased by another person who purchased the book at the same time, or before and after purchasing the book, is presented as reference information. Some virtual bookstores do.

[0003]

When a target book is searched by giving a keyword in a virtual bookstore, usually, the keyword is completely matched with the title without considering the contents of the book. Many books in the field are also hits, and it is difficult to reach the target book from the large number stored in the database. In addition, when a search keyword is input, a word attached to the title of the target book may not be conceived, and thus it may not be easy to reach the target book. In that case, category search is also prepared, but this classification method also contains the subjectivity of the person who performed the classification, so if it does not match your thoughts, you can not reach the target book .

[0004] When a book is presented as reference information by another person at the same time, or before or after the purchase of the book, it is not recommended that the user does not purchase the book. In addition, since no recommendation is made in consideration of the content, it is not always possible to make an appropriate recommendation because it is not possible to distinguish whether the content was purchased at the same time because the content is related or whether the content was purchased simultaneously for another reason. Also,
Since the user's interests, interests, and preferences are not grasped, it does not always work to vaguely search for a book based on the user's own interests, such as searching for a book in a bookstore.

[0005] In the currently established virtual bookstore, the interests, interests, preferences, and hobbies of the user (hereinafter collectively referred to as "preferences") are based on the detailed book description information viewed by the user. There is no one that automatically judges and provides books that are deemed suitable for the user.

The present invention has been made in view of such problems, and automatically determines the user's preference from the detailed book description information and other interactive inputs browsed by the user.
An object of the present invention is to provide a book introduction device capable of efficiently providing a book that the user may be interested in based on the user's preference.

[0007]

Therefore, according to the present invention, as a first aspect, there is provided a user information database for storing and managing user's personal information and preference information, and a book information database for storing and managing book information. , Interactive processing control means for accepting a user's search input or a search request from the user's terminal device, activating necessary processing means and displaying a result or replying to the user's terminal device, and referring to user's preference information Book recommendation means for searching the book information database and recommending and presenting books that match the user's preference, and user profile updating means for updating the user's preference information data from the user's interactive input. It is another object of the present invention to solve the above-mentioned problems by a book introduction device characterized by constantly grasping the user's preference and recommending a book in consideration of the user's preference. To.

According to a second aspect of the present invention, in the book introduction device according to the first aspect, the book information database always stores an introduction document and a certain number of introduction documents for one book to be managed. The user information database includes document vectors in which the number of important words that does not exceed the number of important words are arranged in the order of importance, which is an attribute of the important words, and the user information database includes one or more user profiles for each user. This user profile is a list of words that does not exceed a certain fixed number of words, and each word has as its attribute a numerical value called a learning degree.

[0009] By defining a document vector expressing the contents of a book and a user profile expressing the user's preferences as described above, the quantitative vector necessary to search for a book that matches the user's preferences is obtained. Handling becomes possible.

[0010] In a third aspect of the present invention, in the book introduction device according to the second aspect, the book recommendation unit compares a user profile of a user with a document vector of an introduction document of each book. Then, it has a function of presenting a book with a high degree of coincidence to a user as a recommended book. Here, the degree of coincidence between the user profile and the document vector is the sum of the products of the degrees of importance and the learning degrees for all the words that are commonly present in the document vector and the user profile.

According to a fourth aspect of the present invention, in the book introduction device according to the second aspect, the book recommendation means includes: a document vector of an introduction document of one book selected by a user; The book has a function of comparing with the document vector of the introductory document of each book and presenting it to the user as a related book of the selected book having a high degree of matching. Here, the degree of coincidence between the document vectors is the sum of the products of the degrees of importance for all the words that are commonly present in both. Or, among the books with a high degree of matching,
Furthermore, the document vector of those introduction documents may be compared with the user profile of the user, and a document having a high degree of matching may be presented to the user as a related book of the selected book.

According to a fifth aspect of the present invention, in the book introduction device according to the second aspect, the user profile updating means performs a search according to a user's input of a search word / search sentence or a search genre. When the word is set,
When the document vector of the introductory document of the book selected by the user for the detailed display request or purchase is specified, and / or the words included in the target user profile are necessary at one or both timings. , The operation of updating the target user profile by changing the values of the learning degrees by a predetermined numerical value. The book introduction device configured in this manner constantly keeps track of the latest preferences of the user, so that it follows the changes in the preferences of the user,
Books that match the user's preferences can always be properly introduced.

[0013]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is an overall configuration diagram of a book introduction device 1 according to a preferred embodiment of the present invention. Reference numeral 10 denotes a server computer main body. The server computer main body 10 is connected to a user computer 90 which is a terminal via the Internet 9. A hard disk is connected or built into the computer main body 10, and a book information database 21 and a user profile database 22 are recorded therein.
These databases will be described later. In addition, the computer main body 10 receives a dialogue request from the user terminal device 90, starts necessary search processing, etc., and transmits the result to the user as dialogue screen data to introduce a book to the user. Dialog processing control means 17 for controlling the interactive processing for performing a search, natural sentence analyzing means 13 for extracting a search word from a natural sentence inputted to search for a book, and book recommendation for selecting a book in consideration of user's preference. Means 14, a book database management means 16 for executing search / update / new registration / deletion of the book information database, and a user profile update means 15 for updating a user profile representing user's preference in response to a user's interactive input. I have. The interactive processing control means 17 is, for example, WWW
(World Wide Web) server software and some interactive processing screens in HTML (HyperText Markup Languag)
This can be realized by preparing as a file described in e). HTML files can be stored in other HTML
Set hyperlinks to jump to files,
This is because it is possible to provide a tag that calls a program that executes a specific process. The natural sentence analyzing means 13 is a program module for decomposing sentences given by a known morphological analysis method into words and extracting search words. The book recommending means 14 and the user profile updating means (hereinafter referred to as UP updating means) 15 are program modules called from the interactive processing controlling means 17 as needed. In the following description, the user profile may be abbreviated as UP.

The book information database 21 stores a book master 210, introduction document data 211 of each book, and a document vector 212 of each introduction document data. The book master 210 is provided for each of the books handled by the book introduction device 1.
This is data in which an ISDB (International Standard Book Number) number, classification, book name, author, publisher, selling price, issue year, introduction document data file name,... Are recorded for each book. The introduction document data 211 is a text data file composed of a description of the contents of one book handled by the book introduction device 1 and a book review. The document vector 212 is information that numerically represents the contents of one piece of introduction document data 211.

The document vector will be described. The document vector of a certain document D ₁ is, for example, D ₁ = (Genetic (0.6),
It is expressed as genome (0.5), DNA (0.3), ...). Here, “genetic”, “genome”, “DNA”, and the like are words representing the contents of the document, and the number in parentheses indicates the weight (importance) of the word. The number of important words, that is, the number of dimensions of the document vector is determined as L (for example, L = 10). The words contained in the document are decomposed into individual parts of speech by morphological analysis, which is a known technique,
Obtained by cutting out nouns and proper nouns. The importance of the terms (nouns, proper nouns) obtained in this way can be calculated, for example, using a known technique, TF / IDF method, with respect to a certain document vector Di expressed by Expression 1 as shown in Expression 2. It is determined by the formula. After all, the words obtained as a result of the morphological analysis of the target document are weighted by Equation 2, sorted in descending order of importance, and the document vector is determined by the top L words.

(Equation 1)

(Equation 2)

Reference numeral 22 denotes a user information database.
The user information database 22 records and manages a user master 220 that records registration information of each user and a user profile 222 that is formed as the user uses the book introduction device 1 and that quantitatively expresses a user's preference. I do. User profile 22
2 is formed one or more per user. The user profile can be formally expressed as Equation 3. This is a list composed of several words indicating the user's preference tendency, and each word has a value of learning degree. Like the document vector, it can be treated as a vector quantity.

(Equation 3)

FIG. 2 is a diagram for explaining the outline of the interactive processing of the book introduction device 1. Hereinafter, the book introduction device 2 using FIG.
And the operation of the dialog processing control means 17 will be described.
First, the interactive processing control unit 17 sends the data of the initial screen 100 to the user terminal 90. An initial screen 100 shown in FIG. 11 is displayed on the display of the user terminal 90. On this screen, the user can start searching for a book desired by any method of recommendation search, keyword search, genre search, and natural sentence search.

When the user inputs a search keyword in 112 and clicks the search start button 102 (FIG. 2), the interactive processing control means 17 gives the input search word to the book database management means 16. To search for the relevant book, and the search result is displayed on the search result list screen 20.
0 is displayed as shown in FIG. In 201, the search term used for the search is displayed. Alternatively, when the user inputs a natural sentence to 114 and clicks the search start button 104 (FIG. 2), the interactive processing control unit 1
7 calls the natural sentence analyzing means 13 to extract a search word by a known morphological analysis method. Then, the search term is given to the book database management means 16 to search the corresponding book by giving the extracted search word, and the search result is displayed on the search result list screen 2.
00 is displayed as shown in FIG. 201
The extracted search term is displayed. When the title of any book in the search result book name list 202 is selected (FIG. 2), a detailed explanation screen 300 of the selected book is displayed as shown in FIG.

When the user designates a genre in the list box 113 (FIG. 2), the display of a screen for selecting a hierarchy of several genres is repeated, and then,
Books that belong to the lowest category are identified by your personal information,
The search result list screen 200 is displayed in the order in which the interest and the interest of the user seem to be strong in consideration of the preference information.

When the user clicks the search start button 101 (FIG. 2), the book recommending means 14 considers the user's personal information and preference information, and the book which the user seems to be interested in is likely to be strong. And displays the result as a recommended search result screen 400 shown in FIG. When the title of any book in the search result book name list 401 is selected (FIG. 2)
), A detailed explanation screen 300 of the selected book is displayed.

FIG. 14 shows a detailed explanation screen 300. 3
01 shows the contents of the introduction document of the book. Also,
Reference numeral 302 denotes a title list of books related to the book selected by the book recommendation unit. If the user decides to purchase the book described in detail on this screen 300, he or she clicks the Add to Cart button 303 to add the book (of FIG. 2) to the cart (shopping basket). Then, the display on the display screen shifts to the confirmation screen 500. When the user shifts his or her interest to the related book displayed at 302 on the screen 300, the user selects the title of the related book (FIG. 2), and this time, the detailed explanation screen 300 of the related book is displayed.
Is displayed. If you want to redo the search, click button 304
Click.

The flow of the interactive processing of the book introduction device 1 has been described above. In addition, the dialog processing control means 17 calls the UP updating means 15 at the timing of each user's dialog input / selection in FIG. 2, which will be described later.

Next, the operation of the book recommendation means 14 will be described.
The book recommendation means 14 is a recommended search process which is called when the user instructs the optional search, and a screen 30
A related book search process for displaying 0 is performed. Each process will be described below.

First, the recommendation search process will be described with reference to the flowchart of FIG. First, the book recommendation means 14
Arbitrarily selects one of the user profiles held by the user (S60). Then, the degree of coincidence with the user profile selected in step S60 is calculated from books not purchased by the user managed by the book information database 21 (S63). The calculation of the degree of coincidence is calculated by Expression 4 from the document vector Dq of the introduction document of the book and the selected user profile Pp. Here, Σ is the sum of Tpi = Tqj, that is, the sum of the products of the importance and the learning degree of words that are commonly present in the document vector Dq and the user profile Pp.

(Equation 4)

If the degree of coincidence is the same, and if the degree of coincidence is the same, data is created in which the books are sorted in descending order of publication year, and the result is passed to the dialog introduction means 17 (S6).
6). As a result, the recommended search results are displayed as in a screen 400.

Next, the related book search processing of the book recommendation means 14 will be described with reference to the flowchart of FIG. First, the book recommendation unit 14 calculates the degree of coincidence with the selected book using Equation 5 for all books in the database that the user has not purchased (S70). The degree of coincidence is calculated from the document vector Dp of the introductory document of the selected book and the document vector Dq of the other book by Tpi = Tqj, that is, for words that are commonly present in both document vectors, Take the sum of the products of Next, for books with a high degree of matching, the degree of matching with the currently set user profile of the user is calculated as
(S73). The results are sorted in the descending order of the degree of coincidence obtained in step S73 and in the order of newest issue year, and the result is passed to the dialog introduction means 17 (S76). As a result, the related book list is displayed on the screen 300 as shown at 302.

(Equation 5)

Next, the operation of the UP updating means 15 will be described. The UP updating means 15 performs an UP updating process 900 based on the keyword / natural sentence called at the timing shown in FIG. 2, or an UP updating process 910 based on the selected book called at the timing shown in FIG. 2, or the timing shown in FIG. , UP update processing 930 based on the selected related book called at the timing of FIG. 2, UP update processing 940 based on the selected genre called at the timing of FIG. , And a program module for executing an UP learning degree update process for updating the actual learning degree of the word of the UP. The details will be described below with reference to the flowcharts of FIGS. 3 to 7 and FIG.

FIG. 3 is a flowchart for explaining the UP update process 900 using keywords and natural sentences. First,
In step S10, it is checked whether a natural sentence has been input, and if so, the natural sentence analyzing means 13 is activated, the input is morphologically analyzed, and a search keyword is extracted (S12). Next, some U that the user has
Among the Ps, an UP including all the given search keywords (or extracted in S12) is set as an update target UP (S14). Then, for the update target UP, the value of the learning degree of the word that matches the search keyword is changed to a value obtained by adding 2. However, when those words are connected not by AND but by OR, the value of the learning degree is changed to a value obtained by adding 1 (S16).

FIG. 4 is a flowchart illustrating the UP update process 910 for the selected book. First, by using Equation 4, the user having the highest matching degree with the document vector of the introduction document of the selected book (the introduction document vector of the selected book) is selected from the UPs owned by the user, UP (S20). Next, the introductory document vector of the selected book is set to the attention vector, and the learning degree update value is set to +7.
The P learning degree update process is called (S23).

The operation of the UP learning degree update processing will be described with reference to FIG. First, in step S92, it is checked how many new words exist. Here, the new word is a word that is not included in the target UP but is included only in the attention vector.
When the number of new words is zero or the learning degree update value is negative, no word replacement occurs, but the learning degree update value is positive and the number of new words is one.
If there is more than one,
An upper word having a higher importance in the attention vector is replaced with a lower word having a lower learning degree in the UP (S94). Then, a value obtained by adding the specified update value to the learning degree value of the attention word (word existing in the attention vector) in the UP after the replacement is set as a new learning degree value (S96).

Returning to the description of FIG. In step S23, the learning degree of UP is updated by the selected book introduction vector, and then the learning degree of UP is updated by the book introduction vector of the book not selected. That is, the UP learning degree update process is called with the introduction document vector of the book other than the selected book displayed on 202 of the search result list screen 200 as the attention vector and the learning degree update value = −2 (S2).
6).

This is the end of the description of the UP update process 910 for the selected book. Next, a description will be given of the UP update process 920 using the book placed in the cart with reference to FIG. The books you put in the cart are not the final decision, but you can say that they are very interested in those who decide to purchase at that time,
The UP learning degree update process is called by setting the attention vector as the introduction document vector of the book and the learning degree update value = + 10 (S3).
0).

Next, the UP update processing 930 using the selected related book will be described with reference to FIG. First, the UP learning degree update process is called with the introduction document vector of the unselected book (the book selected in FIG. 2) as the attention vector and the learning degree update value as -3 (S40). The target UP at this time is the UP set in step S20. Next, by using Equation 4, the user having the highest matching degree with the introduction document vector of the selected related book is selected from the UPs owned by the user,
This is set as the target UP (S43). The new target UP is
In many cases, the target UP will be the same as the UP set in step S20, but the target UP may switch here. Finally, the UP learning degree update process is called with the introduction document vector of the related book set to the attention vector and the learning degree update value set to +8 (S46).

Next, the UP update processing 940 according to the selected genre will be described with reference to the flowchart of FIG. First, from the field-specific dictionaries of the selected genre, one of the users having the highest matching degree is selected using Equation 4 (S
50). Here, the genre-specific dictionary is a document vector composed of a set of words that best represent the genre. Next, an UP learning degree update process is called with the dictionary for each field of the selected genre set as the attention vector and the learning degree update value = + 3 (S53).

The book introduction device 1 according to the preferred embodiment of the present invention has been described above. In the description of the UP updating means 15,
Although the timing at which the UP update occurs and the learning degree update value at that time have been specifically shown and described, the gist of the present invention is not limited to the example of this embodiment. In one use of the book use device 1, the UP needs to be updated at least once in response to one or more meaningful interactive operations performed by the user to search for a desired book.

[0036]

As described above, book recommendation means and U
According to the book introduction device of the present invention including the P updating means, the interest, interest, preference, and hobby of the user are automatically determined from the detailed book description information and other interactive inputs viewed by the user. To provide a book introduction device capable of efficiently following a change in a user's preference and efficiently providing a book that the user may be interested in based on the user's current preference. The remarkable effect that it can perform can be produced.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram of a book introduction device 1 according to a preferred embodiment of the present invention.

FIG. 2 is a flowchart of an interactive screen presented by the book introduction device 1;

FIG. 3 is a flowchart illustrating an UP update process using a keyword and a natural sentence.

FIG. 4 is a flowchart illustrating an UP update process using a selected book;

FIG. 5 is a flowchart illustrating an UP update process using a book put in a cart.

FIG. 6 is a flowchart illustrating an UP update process using a selected related book.

FIG. 7 is a flowchart illustrating an UP update process according to a selected genre.

FIG. 8 is a flowchart illustrating a recommended search process.

FIG. 9 is a flowchart illustrating a related book search process.

FIG. 10 is a flowchart illustrating an UP learning degree update process;

FIG. 11 is an explanatory diagram of an initial screen 100 presented by the book introduction device 1;

FIG. 12 is an explanatory diagram of a search result list screen 200 presented by the book introduction device 1;

FIG. 13 is a recommended search result display screen 400 presented by the book introduction device 1.

FIG. 14 is a detailed explanation screen 30 presented by the book introduction device 1.
Illustration of 0

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Book introduction apparatus 9 Internet 10 Server computer main body 13 Natural sentence analysis means 14 Book recommendation means 15 UP update means 16 Book information database management means 17 Interactive processing control means 21 Book information database 22 User information database 211 Introduction document data 212 Introduction document Document vector data of data 222 User profile data 100 Initial screen 200 Search result display screen 300 Detailed explanation screen 400 Recommended search result display screen

Claims (5)

[Claims] 1. A user information database that stores and manages personal information and preference information of a user, a book information database that stores and manages book information, and a processing unit required to receive a search request from a user terminal device. Interactive processing control means for activating and returning a result to the user's terminal; book recommendation means for recommending and presenting a book that matches the user's preference by searching the book information database while referring to the user's preference information User profile updating means for updating the user's preference information data from the user's dialogue input, and constantly grasping the user's preference and recommending books in consideration of the user's preference. Book introduction device. 2. The book information database according to claim 1,
The book contains a document vector in which an introductory document and a number of important words that do not exceed a certain number are arranged in order of importance, and the user information database includes at least one document for each user. The user profile describes user preference information, and this user profile is a list of words that do not exceed a certain fixed number, and each word has a quantity value called learning degree as its attribute. The book introduction device according to claim 1, wherein the book introduction device is provided. 3. The book recommendation means has a function of comparing a user profile of a user with a document vector of an introductory document of each book, and presenting a document having a high degree of coincidence to the user as a recommended book. The book introduction device according to claim 2, wherein the book introduction device is a device. 4. The book recommendation means according to claim 1, wherein the user selects
A document with the function of comparing the document vector of the introduction document of the point book with the document vector of the introduction document of each other book and presenting it to the user as a related book of the selected book with a high degree of matching The book introduction device according to claim 2, which is: 5. The user profile updating means, when a search word is set according to a user's input of a search word / search sentence or a search genre, the user selects a detail display request or purchase. When the document vector of the book's introduction document is specified, at one or both timings, the words included in the target user profile are replaced, if necessary, and their learning level values are determined. 3. The book introduction device according to claim 2, wherein the target user profile is updated by changing only the numerical value, and the user's preference is constantly grasped.