JP2000339322A - Method and device for specifying user attribute and recording medium with its method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To estimate a user's attribute based on the user's behavior having a purpose clearly without the direct input of the user's attribute from a keyboard, etc., by each user. SOLUTION: This device has a user identifier generating part 2 which generates a user identifier for identifying an accessing user, a user identifier transmitting part 3 which transmits the generated user identifier to the accessing user, a user identifier receiving part 4 that receives a user identifier transmitted from the terminal of the user when the user starts a behavior, a macro trend extracting part 5 which extracts a macro trend being the trend of a macroscopic user behavior based on the received user identifier, and a micro user attribute estimating part 6 which estimates the region, racial characteristics, sex, interest and income to which the user having at least a user identifier belongs based on the user identifier, the macro trend, the regional characteristics and temporal characteristics of the macro trend.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a user attribute specifying method and a user attribute specifying device for automatically specifying a user attribute from the behavior of the user, and a recording medium thereof.

[0002]

2. Description of the Related Art Conventionally, the attributes of a user are estimated based on information input directly from a keyboard or the like by each user, or the user's line of sight and movement. It was specified by doing. However, in the method of automatically estimating a user's attribute from the user's gaze or behavior, the user's gaze or behavior is often vague or not clearly aimed, and accurate estimation of the user attribute is difficult. was difficult.

[0003] In the Internet, when collecting attribute information of browser users, a cookie is used as a user identifier for identifying a user. This means that a certain UR
When trying to access L, the HTTP of the URL
The server transmits a user identifier called a cookie to the terminal that has accessed the server. This cookie is not exactly an identifier of each user, but a user identifier given when a certain user uses a certain terminal.For example, if different terminals are used at home and work, They are not necessarily the same user identifier.

[0004] Further, a cookie is for specifying a person who operated a terminal in which the cookie is stored.
When one terminal is used by a plurality of persons, it is difficult to collect a micro user log accurately. Also, if you connect to a different HTTP server with just a cookie, you will have a different server cook
In other words, there is a drawback in that it is necessary to obtain ie, and it is not possible to obtain the true user attributes.
Therefore, the side that opened the home page could not grasp the attribute of the user who actually browsed.

[0005] It is an object of the present invention to select a user's clearly intended behavior without directly inputting the user's attributes using a keyboard or the like, and to select the clearly intended behavior. It is an object of the present invention to provide a user attribute specifying method and a user attribute specifying device for estimating a user attribute based on a user attribute and a recording medium thereof.

[0006]

According to a first aspect of the present invention, there is provided a user attribute specifying method for specifying an attribute of a user accessing via a network, wherein a user identifier for identifying the accessing user is provided. A first step of generating; a second step of transmitting the generated user identifier to the accessing user; and a step of generating a user identifier transmitted from a terminal of the user when the user takes an action. A third step of receiving, and a fourth step of extracting a macro trend, which is a macroscopic user's behavior tendency, based on the received user identifier.
Estimating at least the area, ethnicity, gender, hobby, and income to which the user having the user identifier belongs, based on the user identifier, the macro trend, and the regional and temporal characteristics of the macro trend. And a fifth step.

According to a second aspect of the present invention, in the user attribute specifying method according to the first aspect, a macro trend having the same preference among the macro trends extracted in the fourth step is linked. Characterized by having the following steps:

According to a third aspect of the present invention, in the user attribute specifying method according to the first or second aspect, an identifier provided to the user by the server is collected from a server connected to the network, There is provided a seventh step of tabulating each user together with the user identifier generated in the first step.

According to a fourth aspect of the present invention, in a user attribute specifying device for specifying an attribute of a user who has accessed via a network, a user identifier generating unit for generating a user identifier for identifying the accessing user. A user identifier transmitting unit that transmits the generated user identifier to the accessing user; and a user identifier receiving unit that receives a user identifier transmitted from a terminal of the user when the user takes an action. When,
A macro trend extraction unit that extracts a macro trend that is a macroscopic user's behavior tendency based on the received user identifier, the user identifier, the macro trend, and a regional characteristic and a temporal characteristic of the macro trend. And a user attribute estimating unit for estimating at least the area, ethnicity, gender, hobby, and income to which the user having the user identifier belongs.

[0010] According to a fifth aspect of the present invention, in the user attribute specifying device according to the fourth aspect, a macro that links macrotrends having the same preference among the macrotrends extracted by the macrotrend extracting unit. It is characterized by having a trend linking unit.

According to a sixth aspect of the present invention, in the user attribute specifying device according to the fourth or fifth aspect, an identifier provided by the server to the user is collected from a server connected to the network, It is characterized in that it has a user identifier collection unit that makes a table for each user together with the user identifier generated by the user identifier generation unit.

According to a seventh aspect of the present invention, there is provided a computer-readable recording medium recording a user attribute specifying program for specifying an attribute of a user who has accessed via a network, and identifying the accessing user. A first step of generating a user identifier for the user, a second step of transmitting the generated user identifier to the accessing user, and transmitting the generated user identifier from a terminal of the user when the user takes an action. A third step of receiving an incoming user identifier; a fourth step of extracting a macro trend, which is a macroscopic user's behavior tendency, based on the received user identifier; , Based on the regional and temporal characteristics of the macro trend, at least the user identifier One area to which the user belongs, ethnicity, gender, hobbies, and characterized by recording user attribute specific program and a fifth step of estimating the income.

According to an eighth aspect of the present invention, there is provided a computer readable recording medium storing the user attribute specifying program according to the seventh aspect, wherein the user attribute specifying program is a macro trend extracted in the fourth step. And a sixth step of linking macro trends having the same preference.

According to a ninth aspect of the present invention, there is provided a computer readable recording medium storing the user attribute specifying program according to the seventh or eighth aspect, wherein the user attribute specifying program is transmitted from a server connected to the network. And a seventh step in which the server collects identifiers provided to the users and tabulates them for each user together with the user identifiers generated in the first step.

According to the present invention, by analyzing an access log of a homepage, it is inferred from a user's access behavior what kind of information the user is interested in and what kind of information he wants. (User profile).

Specifically, on a homepage of a search engine (also referred to as a search site) existing on the Internet, more detailed contents of a search keyword entered by a user and a plurality of search result candidates displayed are displayed. It is estimated that, among the homepages selected by clicking for display, the homepage that has taken longer to return to the search result candidate display screen again has a higher degree of interest of the user.

[0017]

Next, an embodiment of a user attribute specifying device according to the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the user attribute specifying device of the present embodiment. It should be noted that the user attribute specifying device of the present embodiment has established a homepage (especially a homepage using a search engine or the like) on the Internet, and specifies an attribute of a user who uses this homepage.

In FIG. 1, reference numeral 1 denotes an interface unit connected to the Internet, and exchanges various information with a user who has accessed the user attribute specifying apparatus of the present embodiment via the Internet. A user identifier generation unit 2 generates an identifier (hereinafter, referred to as a user identifier) for identifying an individual user corresponding to the user who has accessed the user attribute specifying device of the present embodiment. Reference numeral 3 denotes a user identifier transmitting unit that transmits the user identifier generated by the user identifier generating unit 1 to the user corresponding to the user identifier via the user interface unit 1.

Reference numeral 4 denotes a user identifier receiving unit, which receives an identifier of each user via the user interface unit 1 when each user takes an action. Reference numeral 5 denotes a macro trend extracting unit, which macroscopically estimates the tendency of the user's action based on the user identifier received by the user identifier receiving unit 4. Here, the behavior of the user means, for example, that the user instructs the browser to display a certain URL, instructs a homepage search, or requests the server to view a certain content.

Reference numeral 6 denotes a micro-user attribute estimating unit, which includes a user identifier received by the user identifier receiving unit 4 and a macroscopic user behavior tendency estimated by the macro trend extracting unit 5 (representative of a large-scale user set). Based on the behavior (hereinafter referred to as a macro trend) and the regional characteristics and temporal characteristics of the macro trend, the region, ethnicity, gender, hobby, income, etc. to which the user having the user identifier belongs are estimated.
Reference numeral 7 denotes a macro trend linking unit, which links macro trends having the same preference. Reference numeral 8 denotes a user identifier collection unit that extracts a key of the user identifier from each site and collects an identifier of each user based on the key.

9 is a micro user action log storage unit,
It stores an action log of a microscopic user of an individual or of a group having a specific user identifier.
Reference numeral 10 denotes a macro user action log storage unit, which stores a representative action log of a macroscopically large user set, not an individual person. That is, for example, when a certain user goes to another page following a certain URL, the URL followed by the user for each user action is stored in the macro user action log storage unit 10 as a macro user action log. The URL followed by the user is stored in the micro user action log storage unit 9 as a micro user action log in association with the user identifier of the user.

With this, the macro user action log storage unit 10 stores in which URL each user has visited another site and jumped to another site, and also has another URL in the same site. It records whether the page jumped. In addition, the micro user action log storage unit 9 stores
For each user identifier of each user, which URL
Records whether the user visited another site and jumped to another site, or jumped to another page within the same site. Reference numeral 11 denotes a user identifier table storage unit, which stores a table in which all user identifiers of each user are collected.

Next, an operation for specifying a user attribute with the above-described configuration will be described in detail with reference to a flowchart shown in FIG. When the user accesses the user attribute specifying device to search for a desired home page, first, in step S1, the user identifier generation unit 2 uses the same mechanism as an Internet cookie to access the terminal (hereinafter simply referred to as the terminal). Terminal)). For example, when the variable name indicating the user identifier is Key, the user identifier is Key = 123456789.

Here, the variable name Key (hereinafter simply referred to as Key) has the same function as the key of the relational database (hereinafter referred to as RDB) when the user identifier is taken out from the terminal. Also, in order to generate a unique user identifier, a unique user identifier is generated using a function such as an RDB. Also, a user identifier generation unit 2
Causes the generated user identifier to be stored in the microuser action log storage unit 9.

Next, proceeding to step S2, the user identifier transmitting section 3 first inquires of the terminal whether there is a user identifier registered with the key using the key. If there is a response that there is no user identifier registered with Key, the process proceeds to step S3,
The user identifier generated by the user identifier generation unit 1 is transmitted to the user side and stored in the storage device of the terminal. Also,
If a user identifier registered with Key already exists, the process proceeds to step S4, and the terminal is transmitted with the user identifier.

Here, in the above operation, first, in step S1, after generating a user identifier for the terminal,
In step S2, an inquiry is made to the terminal as to whether there is a user identifier. First, an inquiry is made to the terminal as to whether there is a user identifier.
When a response indicating that there is no user identifier is received, a user identifier may be generated and stored in the storage device of the terminal.

When the processing in step S3 or S4 is completed, the process proceeds to step S5, where the user identifier receiving unit 4
Receives a user identifier of a user transmitted from a terminal used by the user when the user takes an action. Here, the actions taken by the user include:
The act of inputting a search key for a homepage search, the result of a homepage search, the homepage title or URL (Uniform Resource) displayed on the user terminal screen.
Locator), in order to jump to a page desired by the user, there is an act of designating the title or URL of the page.

Next, proceeding to step S6, the macro trend extracting section 5 and the micro user attribute estimating section 6 extract and extract a macro trend based on the user identifier.
First, the extraction of the macro trend by the macro trend extraction unit 5 will be described. The macro trend extraction unit 5 extracts a macro trend by accumulating and analyzing a macro user action log. This is described in, for example, "Information associating apparatus and method" by Okubo et al.
The macro trend can be extracted with high accuracy by using -320419.

Here, the procedure of extracting a macro trend executed by the macro trend extracting unit 5 will be described with reference to FIG. First, the macro trend extraction unit 5 determines whether or not a search key for searching for a desired homepage has been input by the user via the user identifier reception unit 4 (step Sa1). If the search key has not been input, the determination result is NO, and the macro trend extraction unit 5 ends the process.

If a search key is entered, the result of the determination is YES, and a search is made for a home page based on the input search key. The title or URL of the searched home page and its contents ( That is, the search result) is displayed in a list on the user terminal (step Sa2). At the same time, timing is started, and it is determined whether or not the user has clicked on any of the titles or URLs of the pages listed as search results within a predetermined time (for example, 1 to 2 minutes) (step). Sa3).

Here, if the title or URL of the page displayed in the list is not clicked at all within the predetermined time, the determination result is NO, and the macro trend extracting unit 5 ends the processing. If any title or URL is clicked within the predetermined time, the determination result is YES, and the clicked title or URL (search result) has been clicked by a user having the same user identifier in the past. It is determined whether or not it has been performed (step Sa4).

If the same title or URL clicked in step Sa3 has been clicked by a user having the same user identifier in the past, the determination result is YES, and the macro trend extraction unit 5 The process ends. On the other hand, if the title or URL of the page clicked by the user for the first time is NO, the search key input by the user in step Sa1 and the title or URL of the page clicked by the user And the total number of times the clicked page has been clicked so far (that is, including the number of times clicked by another user) is stored in the macro user action log storage unit 10 as a macro user action log. Remember. Here, it can be considered that the total indicates the magnitude of the tendency (trend strength).

Next, the macro trend extraction unit 5 refers to the macro user action log stored in the macro user action log storage unit 10 and determines whether the value of Δcumulative / Δtime is equal to or more than a certain value. Is determined (step Sa6). Here, the value of [Delta] total / [Delta] time indicates the degree of the total (the slope of the trend) that changes over time. If the value of Δcumulative / Δtime is less than the fixed value, the determination result is NO, and the macro trend extracting unit 5 ends the process. If the value of Δcumulative / Δtime is equal to or greater than a certain value, the determination result is YES, and after a predetermined time has elapsed, whether or not the value of Δcumulative / Δtime is equal to or less than a certain value. Is determined (step Sa7).

If the value of .DELTA.cumulative / .DELTA.time exceeds a certain value, it is considered that the cumulative value has not yet reached the peak, the determination result is NO, and the macro trend extracting unit 5 terminates the processing. . On the other hand, if the value of Δtotal / Δtime after the lapse of the predetermined time is equal to or less than the predetermined value, it is considered that the cumulative total has reached the peak, the determination result is YES, and the next step S
It is determined whether or not the total number of clicks of the page clicked in a3 is equal to or more than a certain value (step Sa).
8). Here, if the total is less than the predetermined value, the determination result is NO, and the macro trend extraction unit 5 determines that the content of the search key and the search result are not the macro trend, and ends the processing.

On the other hand, if the total is equal to or more than a certain value,
The result of the determination is YES, and the search key input in step Sa1, the search result (page title or URL) clicked in step Sa3, and the time when the cumulative click on the page reaches a peak are defined as a macro trend. The data is stored (step Sa9), and the apparatus is again in a standby state for inputting a search key (step Sa1).

By performing the above-described processing, for example, during the time when the cherry blossoms bloom, the search for “sakura” and the “sakura”
The number of jumps to the sites related to 増 加 increases, and a macro trend (a general trend) in which a peak appears at the end of March, which is the time when cherry blossoms bloom, as shown in FIG. 4, can be extracted.

Next, estimation of user attributes by the micro user attribute estimation unit 6 will be described. The micro-user attribute estimating unit 6 determines the user identifier based on the user identifier received by the user identifier receiving unit 4, the macro trend extracted by the macro trend extracting unit 5, and the regional characteristics and temporal characteristics of the macro trend. , The ethnicity, gender, hobby, income, etc. of the user having

That is, the micro user attribute estimating unit 6
From the macro trend extracted by the macro trend extraction unit 5, a “trend” based on regional characteristics and temporal characteristics is further extracted. For example, in the case of a "trend" having a regional characteristic, the regional axis is spread and a macro trend is calculated. If the trend is based on temporal characteristics, the time axis is spread and the macro trend is recalculated.

As a result, the macro trend extracted by the macro trend extraction unit 5 changes to a regional trend or a temporal trend. Then, the characteristics of the user can be extracted by dynamically analyzing the regional trends and the temporal trends. For example, when the macro trend shown in FIG. 4 is extracted for each region, some peaks appear as shown in FIG. Note that the dotted line in FIG. 5 indicates the macro trend in FIG.

This indicates that the flowering of "cherry blossoms" is a trend that has an axis both temporally and regionally. That is, in FIG. 5, the first peak is, for example, a macro trend in the Kyushu region, the next peak is a macro trend in Tokyo or Honshu, and the last peak is a macro trend in the Tohoku region or Hokkaido region. .

By extracting the user identifier causing the macro trend from the micro-user action log storage unit 9, the user having a certain user identifier can be identified as a person who lives in the Tokyo area or a user in the Kyushu area. It is possible to estimate a user's local attribute, such as a person who lives or a person who lives in the Tohoku region. When further narrowing down the user's local attribute locally, the regional macro trend as described above is further compounded and narrowed down, for example, a user with a certain user identifier is living in Suginami-ku, Tokyo, It is possible to specify the user attribute of how much the annual income is when living alone.

That is, when a plurality of regional macrotrends are multiplied to form a composite, if each macrotrend is considered to be a set, the sum of the set will be obtained. Can be. For example, if one macro trend identifies users in the Kanto region and another macro trend identifies users north of Tokyo, these two types of macro trends identify users in the Kanto region and north of Tokyo. can do. In this example, it is estimated that the users are located in Tokyo, Saitama, and Ibaraki.

Hereinafter, the procedure for estimating the user attribute in the micro user attribute estimating unit 6 will be described with reference to the flowchart shown in FIG. First, similarly to the macro trend extraction unit 5, the micro user attribute estimation unit 6 determines whether or not a search key for searching for a desired homepage has been input by the user via the user identifier reception unit 4 (step). Sb1). If the search key has not been input, the determination result is NO, and the microuser attribute estimating unit 6 ends the process.

On the other hand, if a search key is entered,
As described above, the homepage is searched by the macro trend extraction unit 5 based on the search key input by the user, and the search results are displayed in a list on the user terminal (see step Sa2 in FIG. 3). At this time, step S
The result of the determination of b1 becomes YES, and a predetermined time (for example, 1
It is determined whether or not the user has clicked on any of the titles or URLs of the pages listed as search results within 2 minutes (step Sb2).

Here, if the title or URL of the page displayed in the list is not clicked at all within the predetermined time, the determination result is NO, and the microuser attribute estimating unit 6 ends the processing. If any title or URL is clicked within the predetermined time, the determination result is YES, and the micro-user action log storage unit 9 is clicked.
The search key input by the user in step Sb1 and the title or URL of the page clicked by the user are associated with the user identifier of the user stored in the micro user action log as a micro user action log. It is stored in the storage unit 9 (step Sb3). Here, it can be considered that the total indicates the magnitude of the tendency (trend strength).

Next, it is determined whether or not the clicked title or URL (search result) matches the macro trend extracted by the macro trend extraction unit 5 (step Sb4). Here, this determination is made when the homepage indicated by the URL of the clicked search result includes a word indicating the macro trend, or when the URL of the clicked search result itself determines the macro trend, If another user is clicking, it is determined that they match.

If the clicked search result does not match the extracted macro trend, the determination result is NO, and the micro user attribute estimating unit 6 ends the processing. On the other hand, if they match, the determination result is YES, and it is determined whether or not the matching macro trend is a trend that specifies the user attribute (step Sb5). Here, the trend for specifying the user attribute can specify the attribute of the user such as the user's age, family structure, residence area, hobbies, preferences, things to purchase, objects of interest, events of interest, and the like. A trend. For example, in a case where the user's residence area is specified, when the user attempts to acquire information related to the user's residence area (for example, flowering time of cherry blossoms near the user's residence area, flowering prediction, etc.), Is specified. In this case, the trend specifying the user attribute (living area) is “cherry blossom blooming time or flowering prediction near the residential area”.

If it is a trend specifying a user attribute, the determination result is YES, and the macro trend is stored as a user attribute (step Sb6). On the other hand, if it is not a macro trend that specifies user attributes,
If the determination is NO, the macro trend is stored as a user attribute together with the user identifier (step S).
b7). By the processing in step Sb6, what has not been determined as a macro trend until now is determined as a new macro trend.

Next, returning to FIG. 2, the process proceeds to step S7, where the macro trend linking section 7 links macro trends having the same preference. This can be related by physically linking the same product or a product expressing a concept in another language based on a macro trend.

The macro trend linking section 7 of this embodiment performs a morphological analysis on the macro trend of the user estimated by the macro trend extracting section 5. Morphological analysis refers to a search for an input character string in a word dictionary, part-of-speech information (part-of-speech), head-of-sentence information (head-of-sentence possible), forward connection information (antecedent), backward connection information (adjoint) And other information. A normal word dictionary has a special dictionary structure called a TREI dictionary structure so that high-speed search can be performed.

For example, when there are dictionary items "Ah", "Aiatsu", and "Ai", each of the first characters (in this case, since it is Japanese, different from the alphabet which is a character of C language, a Japanese character 2 bytes) are the same, the second character is the same, and so on. When the last character is matched, information such as part of speech information (part of speech), sentence head availability information (head part possible), forward connection information (front connection), back connection information (back connection), etc. for the word dictionary item Is described.

Here, the head-of-sentence information is a flag indicating whether or not it can be at the head of the sentence. If the beginning of the sentence is acceptable, the word may be present at the beginning of the sentence, but if the sentence is not the beginning of the sentence, the word is not permitted to be at the beginning of the sentence. The forward connection information is permitted to be connected only when the part of speech or attribute of the previous word is appropriate, and is deleted as a candidate when the word is not permitted to be connected in the preceding word. Similarly, backward connection information is also
Connection is permitted only when the part of speech or attribute of the subsequent word is appropriate, and in the case of a word for which connection is not permitted in the subsequent word, it is deleted as a candidate.

The maximum likelihood candidate is selected by such a part of speech connection. The maximum likelihood candidate is selected by a method called a minimum cost method. The minimum cost method is a processing method in which a morpheme candidate having the lowest cost is set as a maximum likelihood candidate. The costs used in morphological analysis generally include two types of costs: connection costs and word costs.

The connection cost is a cost required to connect a certain word with another word. Since a word is a word, the connection cost for a word obtained by adding the word to the word is “0”. The word cost is a cost related to the word. For example, a frequently used word has a low cost. In addition, since utilization is not a word, the cost is “0”.

By this morphological analysis, the text part is decomposed into words and, at the same time, the parts of speech considered to be most suitable for each word are added. For example, "Preste",
"Farby" is subjected to morphological analysis as shown in FIGS. 7 (a) and 7 (b).

Here, unknown words or proper nouns are extracted from the macro trends extracted by the macro trend extraction unit 5. Excludes other macro trends. Then, a macro trend having a different expression is extracted from the macro trends generated at the same time or slightly shifted in time. Then, when the temporal trend is similar or the trend of the macro trend is similar, the macro trend can be linked by indicating that the macro trend indicates a similar target (for example, a product name or an event). Here, similar in time means that, for example, if a certain product is released simultaneously in Japan and the United States, or a trend that occurred in the United States and Japan on the same day, a similar trend occurs with a time difference from the United States. ,
This is the case.

For example, as shown in FIG. 8, a macro trend in the United States and a macro trend in Japan are compared, and macro expressions having different expressions and similar in trend (change in trend intensity over time) are shown. Assuming that macro trends expressed as “Furby” in the United States and macro trends expressed as “Farby” in Japan are extracted as trends, the target of these macro trends indicates similar products. "Furby
= Farbi ”can link each other's macro trends.

Next, proceeding to step S8, the user identifier collection unit 8 extracts the key of the user identifier from each site and collects each user identifier based on the key. It receives HTTP set_cookie information for each site, extracts the key information at each site, and uses the key information to inform the user who has entered a specific URL site.
By performing get_cookie based on the Key, the key generated by the user identifier generation unit 2 in FIG. 1 and its user identifier correspond to the key provided by another site and its user identifier for each user. And stored in the user identifier table storage unit 11. Then, all the plurality of user identifiers possessed by a certain individual user are output from the user identifier table storage unit 11 to the micro user action log storage unit 9.

For example, when a certain individual user U visits various sites, the user U does not have only a specific user identifier, but a certain A site has “111”.
In some cases, the other site B has a user identifier of “2222.” Therefore, in order to integrate those user identifiers, the user identifier of the user U is “1111”. And "22
22 "is stored in the user identifier table storage 11 and the user identifier table storage 1
1 to the micro-user action log storage unit 9, all the plurality of user identifiers of a certain individual user are output.

For example, the user identifier “1111” of the user U is stored in the micro user action log storage unit 9 as URL: ww
The micro user action log jumping to w.123.co.jp and the user identifier “2222” of the user U are also UR
L: When the micro user action log jumping to www.aaa.co.jp is stored, the two logs are integrated, and the user U makes the URL: www.123.co.jp and UR
L: Specify that two actions of www.aaa.co.jp were performed. As a result, even a user who has another cookie at another site can be centrally managed with the user identifier provided by the user attribute specifying device of the present embodiment.

The interface unit 1 shown in FIG.
User identifier generating unit 2, user identifier transmitting unit 3, user identifier receiving unit 4, macro trend extracting unit 5, micro user attribute estimating unit 6, macro trend linking unit 7, user identifier collecting unit 8, micro user action log storage unit 9, a computer-readable recording of a program for realizing the functions of the macro user action log storage unit 10 and the user identifier table storage unit 11 or the programs for realizing the procedures of FIGS. 2, 3, and 6 Recorded on a medium,
By reading and executing the program recorded on the recording medium by the computer system, the communication rate of the network allocated to the user may be controlled.

Here, the “computer system” includes an OS and hardware such as peripheral devices.
If a WW system is used, a homepage providing environment (or display environment) is also included. The “computer-readable recording medium” is a floppy disk, a magneto-optical disk, a ROM, a CD-R
It refers to a portable medium such as an OM or a storage device such as a hard disk built in a computer system. Further, the “computer-readable recording medium” refers to a communication line for transmitting a program via a network such as the Internet or a communication line such as a telephone line, and dynamically holds the program for a short time. In this case, a medium that holds a program for a certain period of time, such as a volatile memory in a computer system serving as a server or a client in that case, is also included.

The above-mentioned program may be for realizing a part of the above-mentioned functions, and may be for realizing the above-mentioned functions in combination with a program already recorded in the computer system. May be.

[0064]

As described above, according to the present invention,
By automatically estimating the user's attributes based on the intended behavior, the user's physical location,
Attributes that characterize the user, such as status, ethnicity, occupation, gender, and hobbies, can be automatically extracted. Unlike user characteristics intentionally described by the user, the user characteristics under the user's unconsciousness are automatically extracted. It can be collected and used as basic data such as marketing.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of an embodiment of a user attribute specifying device according to the present invention.

FIG. 2 is a flowchart showing an operation of the user attribute specifying device.

FIG. 3 is a flowchart showing a procedure for extracting a macro trend in the user attribute specifying device.

FIG. 4 is a diagram illustrating an example of a macro trend.

FIG. 5 is a diagram showing a case where the macro trend shown in FIG. 4 is extracted separately for each region.

FIG. 6 is a flowchart showing a user attribute specifying procedure in the user attribute specifying device according to the present invention.

FIG. 7 is a diagram showing an example of morphological analysis.

FIG. 8 is an explanatory diagram for explaining linking of a macro trend.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Interface part 2 User identifier generation part 3 User identifier transmission part 4 User identifier reception part 5 Macro trend extraction part 6 Micro user attribute estimation part 7 Macro trend linking part 8 User identifier collection part 9 Micro user action log storage part 10 Macro user Action log storage unit 11 User identifier table storage unit

Continued on the front page F term (reference) 5B049 AA01 EE05 FF01 GG09 5B075 PR03 QM05 5B089 GA11 GB03 HA10 JA21 JA37 KA14 KC48 KC53

Claims (9)

[Claims] 1. A user attribute specifying method for specifying an attribute of a user who has accessed through a network, a first step of generating a user identifier for identifying the accessing user, and the generated user identifier. A second step of transmitting a user identifier transmitted from the terminal of the user when the user takes an action, a second step of transmitting a user identifier transmitted from the terminal of the user, A fourth step of extracting a macro trend which is a macroscopic user's behavior tendency based on the user identifier, the macro trend, and a regional characteristic and a temporal characteristic of the macro trend. Fifth step for estimating the region, ethnicity, gender, hobby, and income to which the user having the identifier belongs. User attribute specifying method characterized in that it comprises and. 2. The method according to claim 1, further comprising the step of linking macrotrends having the same preference among the macrotrends extracted in the fourth step. . 3. A server which collects, from a server connected to the network, the identifier provided by the server to the user, and tabulates the table for each user together with the user identifier generated in the first step. The method according to claim 1, further comprising a step. 4. A user attribute specifying device for specifying an attribute of a user accessing via a network, a user identifier generating unit for generating a user identifier for identifying the accessing user, and the generated user identifier. A user identifier transmitting unit that transmits the user identifier to the accessing user; a user identifier receiving unit that receives a user identifier transmitted from a terminal of the user when the user takes an action; and the received user identifier. A macro trend extraction unit that extracts a macro trend that is a macroscopic user's behavior tendency based on the user identifier, the macro trend, and a regional characteristic and a temporal characteristic of the macro trend. Estimate the area, ethnicity, gender, hobbies, and income of the user with the identifier. User attribute specific apparatus characterized by having the user attribute estimator for. 5. The user attribute according to claim 4, further comprising a macro trend linking unit that links macro trends having the same preference among the macro trends extracted by the macro trend extracting unit. Specific device. 6. A user identifier that collects, from a server connected to the network, an identifier provided by the server to the user, and creates a table for each user together with the user identifier generated by the user identifier generating unit. The user attribute specifying device according to claim 4, further comprising a collection unit. 7. A computer-readable recording medium recording a user attribute specifying program for specifying an attribute of a user who has accessed via a network, wherein a user identifier for identifying the accessing user is generated. A first step, a second step of transmitting the generated user identifier to the accessing user, and receiving a user identifier transmitted from a terminal of the user when the user takes an action. A third step, a fourth step of extracting a macro trend, which is a macroscopic user's behavior tendency, based on the received user identifier; the user identifier, the macro trend, and a regional trend of the macro trend. Based on the characteristics and temporal characteristics, at least the place to which the user having the user identifier belongs , Ethnicity, gender, hobbies, recording medium recording a user attribute specific program and a fifth step of estimating the income. 8. The user attribute specifying program according to claim 7, further comprising a sixth step of linking macro trends having the same preference among the macro trends extracted in the fourth step. A computer-readable recording medium on which is recorded. 9. A seventh method comprising: collecting, from a server connected to the network, an identifier provided by the server to the user, and tabulating the collected identifier with the user identifier generated in the first step for each user. 9. A computer-readable recording medium on which the user attribute specifying program according to claim 7 or 8 is provided.