[go: up one dir, main page]

WO2002001411A1 - Systeme fournissant des informations relatives au sens des valeurs - Google Patents

Systeme fournissant des informations relatives au sens des valeurs Download PDF

Info

Publication number
WO2002001411A1
WO2002001411A1 PCT/JP2001/005530 JP0105530W WO0201411A1 WO 2002001411 A1 WO2002001411 A1 WO 2002001411A1 JP 0105530 W JP0105530 W JP 0105530W WO 0201411 A1 WO0201411 A1 WO 0201411A1
Authority
WO
WIPO (PCT)
Prior art keywords
user
data
value
preference
code
Prior art date
Application number
PCT/JP2001/005530
Other languages
English (en)
Japanese (ja)
Inventor
Shinichi Sakurai
Original Assignee
Shinichi Sakurai
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shinichi Sakurai filed Critical Shinichi Sakurai
Priority to AU2001266343A priority Critical patent/AU2001266343A1/en
Publication of WO2002001411A1 publication Critical patent/WO2002001411A1/fr

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation

Definitions

  • the present invention reduces the burden on users in searching for information on sites (pages) distributed over an information communication network such as the Internet, and facilitates the information of sites that each user wants individually.
  • an information communication network such as the Internet
  • the present invention provides a recommendation (recommended for analysis of preference (value) analysis) based on a swarm intelligence inference technology applied to artificial intelligence (AI) in an information communication network such as the Internet. 1) Technical techniques that enable provision of individual information. Background art
  • the conventional concept of artificial intelligence is to determine all samples from phenomena and phenomena that occurred in the past, create certain rules based on these, and derive future guesses and behavior predictions. .
  • the current artificial intelligence is a stand-alone idea, so it is optimal for a specific closed condition or set.However, it is not a structure that is conscious of the network and unspecified majority, but each one is The system is complex and bloated to pursue events. To predict the movement of an individual, the more elements there are, the closer to the optimal solution, but the more elements there are, conversely, the more complex the system and processing.
  • An object of the present invention is to provide a method that allows any user to easily and freely use an information communication network such as the Internet.
  • Another object of the present invention is to provide a method capable of providing information of a site that each user wants individually by using human-based artificial intelligence (AI).
  • AI artificial intelligence
  • the value information providing system of the present invention uses a “recommended” site discovered by one of the users to determine the value (hobby, taste, personality, etc.) closest to the user. Recommend and display to owners. Perform matching process using artificial intelligence based on values. '
  • the innumerable information of the sites distributed on the information communication network such as the Internet is two-dimensionally (or more) and mainly for humans (human-based). ) Can solve problems that previously required complicated programs and a considerable amount of hardware with simple processing methods.
  • the solution is to use an information transfer method that is used by insects that act in swarms, such as “ari”.
  • swarm intelligence or social insects, teaches Complex systems are applied to the Internet as swarm intelligence estimation technology. It also enables the exchange of user personal preference information with other site servers.
  • a 1 to 1 bookmark (link) collection corresponding to each user is automatically generated as a tool for dynamically capturing each user's values (preferences).
  • the site name of this link collection, the evaluation comment of the site, and the character string of the classification attribute and the flag indicating the content of the commenter are analyzed to grasp the individual user's preference contents, By digitizing or symbolizing (value code Taste Code) with a certain rule, the preference is captured by finding the position of the individual in the group.
  • a value information providing method in which a priority is determined by use frequency and classified, a link collection of display contents specific to a user is generated by linking to an original link collection of a commented data provider;
  • the method of estimating swarm intelligence captures the dynamic change of a bookmark created in real time for each user individually, generates a summary code by character string analysis, and determines the positional relationship between the whole group and each user. By utilizing the information, the user's individual preference is estimated. ... ' BRIEF DESCRIPTION OF THE FIGURES
  • Figure 1 is an explanatory diagram in which two people who searched for the same theme in two different windows on the “recommended” display window and recommended them are displayed as “recommended”.
  • Figure 4 is an explanatory diagram in which the preference attribute is selected from “who” data based on the content registered by the user, and is quantified by the composition ratio of “from” the personal values.
  • Figure 5 shows the dynamic Taste that is closest to the dynamic TasteCode.
  • Figure 6 is an explanatory diagram that analyzes whether a user's preference is composed of “who (plural)” preference, and selects and displays the “recommended” that is optimal for that user.
  • FIG. 10 is an image diagram of the configuration of the agent of the user E.
  • Figure 12 is an explanatory diagram of how to generate absolute TasteCode from user E's theme unit.
  • Figure 13 is an illustration of how to give recommendations to the user E's theme unit. It is.
  • FIG. 14 is an explanatory diagram of a method of conversing using the theme unit of user E.
  • Figure 15 is an explanatory diagram of how to display an advertisement using the theme unit of User E.
  • Fig. 16 is an explanatory diagram of how users with the same preference communicate using the theme unit.
  • Figure 17 is an illustration of the outline, symbols, and abbreviations of the automatic bookmark collection.
  • Figure 18 is an explanatory diagram of the AI site list (menu) screen.
  • FIG. 19 is an explanatory view of the screen of the character “TassteUnit” and the character attributes, “DataCode”, and the CDC.
  • Figure 20 is an explanatory diagram of the automatic link collection screen of user Ux and how to give recommendations based on guess at reduction and enlargement.
  • Figure 21 is a schematic diagram of the actual homepage display from the smallest unit, DataUnit.
  • Figure 22 shows a method and a schematic diagram for estimating the user's interests and preferences by estimating reduction and enlarging.
  • Figure 23 shows a method and a schematic diagram for estimating the user's interests and preferences by reducing and estimating the enlargement.
  • Figure 24 is an explanatory diagram of the absolute values code in the extended speculation of recommendations.
  • Figure 25 is an explanatory diagram of how to obtain and proceed with the data of the reduction estimation.
  • Figure 26 is an explanatory diagram of how to obtain and proceed with the data for enlargement estimation.
  • Figure 27 shows the structure and explanation of the character screen.
  • Figure 28 is an explanatory diagram when there are multiple character attributes.
  • Fig. 29 is a list (menu) screen of AI sites operated during use and an explanatory diagram.
  • C Fig. 30 is a list of AI sites listed (menu) and is an image diagram when introduced to a newspaper company.
  • Figure 31 is an explanatory diagram of the character screen when introduced to a newspaper company.
  • Figure 32 shows the actual home base from the smallest unit, Data Unit, at the AI site.
  • Figure 33 shows the relationship between the character screen and TU (theme unit) when introduced to a newspaper company.
  • Figure 34 is an explanatory diagram of a method of estimating the user's interests and preferences by estimating reduction and enlargement when introduced into a newspaper company.
  • Figure 35 is an explanatory diagram of the method of confirming conversational interest and reflecting it in TasteCode.
  • the server classifies a user (person) as a minimum unit, and human-based artificial intelligence (a program for displaying an optimal “recommended” for the user). Is a search engine that matches people's values (hobbies, preferences, etc.).
  • the server is a value information database server.
  • the server is a server computer and, like the terminal device, consists of a central control unit, a main storage device, a hard disk device as an external storage device, a communication interface, and a display device as a display device. I have.
  • the terminal device includes an input device such as a keyboard and a mouse in addition to the above component devices.
  • no keyboard is used for inputting a keyword for search.
  • the server's central controller operates according to the program installed on the hard disk drive. Various functions of the server are performed by this program.
  • HTTP Hyper Text Transfer Protocol
  • HTML Hyper Text Markup Language
  • HTMLff Dynamic Markup Language
  • XML Extensible Markup Language
  • the central control unit of the terminal device displays the information distributed from the server on the display device according to the program installed in the hard disk device.
  • the central control unit of the terminal device transmits the specified information and commands input through the mouse to the server via the communication interface, communication line, and network.
  • This swarm intelligence AI recommendation allows each user to capture a dynamic change in the puck mark created in real time and analyze the character string to generate a summary code (Taste Code), and the positional relationship between the entire group and each user. This is a method for estimating individual user preferences using
  • UX For example, to capture the taste of user UX, sort this code according to a certain rule, determine the position of UX among all users, and determine U-1 and U-1 before and after Ux. Guess Ux's preferences by comparing it to U + 1's code content and bookmark content.
  • the value information providing system SYS uses a swarm algorithm on the Internet and implements it on a Web site (such as a home page) that uses artificial intelligence and speech recognition.
  • the server 1 in the value information providing system SYS can execute the following processes.
  • Dynamic HTML real-time using various CGIs, etc.
  • dynamically grasping user needs using file such as woodpecker C0 okie
  • the individual content generation usage frequency priority classification link collection in which the content displayed by each user is different is generated (see Fig. 1).
  • Site name and site evaluation comments and classification attributes which are the contents of the individual content generation and usage frequency priority classification link collection in (2) above, which includes this arrangement order (priority of user preference).
  • the user can use artificial intelligence (such as a program that displays the “recommended” best for that user) to display the “recommended” to the user.
  • the basic preference attributes are acquired, and the user's preference priorities are determined and used as basic user preference attribute data (see Fig. 3).
  • a recommended acquisition type individual content generation use frequency priority classification link collection is generated (see Fig. 1).
  • Main server on the Internet generated by XML or dynamic HTML (separate real-time or dynamic programs using various CGIs etc.) that can use data separately. ), And synchronizes the link collection (recommended site information of other people, etc.) on another server in a distributed location with the link collection (recommended site information, etc.) on the main server, and the same data in real time. Acquire and share (see Figure 7).
  • the server 1 that is both a web server and a value information database server responds to a designated operation, such as a mouse click operation, by each user using the terminal device on its own web site. Collect data as value information. Server 1 accumulates the collected value information in its own database as text data (text files).
  • the terminal device 2 When the user using the terminal device 2 accesses the top page of the site provided by the server 1, the terminal device 2 displays the screen shown in FIG. This display screen is provided with a “recommended” display window 10.
  • the recommendation display window 10 shows a push-type basic site where different users (people) have searched for (used once) with the same theme and recommended them. Typically, two items are displayed (up and down side by side).
  • the user ⁇ ⁇ selects one of the two “recommended” displayed in the display window 10 by a designated operation
  • the user ⁇ is displayed as the data unit DU of the link collection LK and is displayed in the database D ⁇ of the server 1. be registered. If a data unit DU corresponding to user ⁇ has already been registered in database D ⁇ of server 1, the selected data unit DU is additionally registered.
  • server 1 guesses the next “recommended” and displays it on display window 10. Note that the above process of generating the link collection LK specific to the user E is the same when registering the link collection LK for another user who has already been registered.
  • the server 1 combines a cookie or identification code, dynamic HTML (also XML, etc.), and a swarm intelligence inference method (artificial intelligence) to use each user. Provide the best sorting directory service in order of frequency (best ranking).
  • the server 1 when the server 1 receives access from the user, the server 1 uses file information such as a cookie that recognizes the user of the web page, or identification code such as an ID or password, and uses dynamic HTML and other various types of information.
  • file information such as a cookie that recognizes the user of the web page, or identification code such as an ID or password, and uses dynamic HTML and other various types of information.
  • a link collection (individual content generation use frequency priority classification link collection) with different contents displayed by each user is generated by a program that captures user needs dynamically and in real time using CGI and the like.
  • the screen of the link collection LK including the data units DU in a completely different order is displayed.
  • the server 1 uses a predetermined character string based on a character string corresponding to information on basic attributes (age, gender, smoking cessation, etc.) of the self-report, which is input when the user E accesses his / her site through the terminal device 2 for the first time. Digitize with rules (predetermined conversion codes such as JIS) and generate, for example, "Taste Code TC I" 1 00 0 1 0 0 1 0 1 0 1 0 1 1 1 1 1 0 1 0001 1 1 1 "for basic attributes I do.
  • rules predetermined conversion codes such as JIS
  • the data unit (minimum unit) DU is displayed on each line of the link collection LK based on the operation specified by the user E for the “recommended” display on the display window 10. For example, the data unit DU specified by user E And gourmet ", comment” Introduces various domestic trips, hot springs and gourmet information is abundant “, and flag” A2 (second recommended by link author A) ".
  • the order of the registered link that is, the data unit DU changes.
  • the order of T a s e C o d e changes dynamically in response to this change.
  • the user's own position and correlation within the entire group (group) can be dynamically changed by using each one, and it is possible to grasp real-time personal preferences (needs). Become. As a result, it becomes possible to grasp the taste of a specific group in real time.
  • the server 1 determines which data unit DU the preference attribute is selected from based on the content registered by the user, and who the personal value is. Is quantified by the composition ratio.
  • users A, B, and C are providers of “recommended” data for user E. T customer / 05530
  • the flag FG in Fig. 4 indicates that the data provider is User A and is the fifth “recommended” of this user.
  • server 1 obtains a dynamic taste C based on the user's personal preference attribute basic data obtained in (3) above and the numerical values obtained in (5) above.
  • server 1 obtains a dynamic taste C based on the user's personal preference attribute basic data obtained in (3) above and the numerical values obtained in (5) above.
  • the components of the dynamic taste code closest to the ode select the “recommended” most suitable for the user. All users are represented by the value of Taste Code, and the “recommended” contents to be displayed to a certain user are also displayed by analyzing the contents (link collection LK) of the user who has the closest Taste Code to this user. Is done. ⁇
  • the server 1 uses artificial intelligence to display to the user “recommended” the content that matches his or her preference from data of people with similar values, centering on himself (a certain user).
  • server 1 determines the most suitable “recommended” for that user. Choose out. In other words, the server 1 analyzes which (multiple) preferences the preference of a certain user is, and selects and displays the “recommended” for the user.
  • the value composition ratio of user E is 30% for user A, 15% for user B, 10% for user C, and 45% for other users.
  • it consists of
  • server 1 first displays “recommended” (data unit DU) from “recommended” (link collection LK) provided by user A with a high composition ratio on terminal device 2 of user E, If E is a pass (if you are not interested), the following configuration “Recommended” from User B and User C with high elements are simultaneously displayed on “Recommended” display window
  • the display window 10 may be described as a display small window 1 ⁇ .
  • the server 1 displays the “recommended” of the user B. If the number of components exceeds 50% consecutively, Server 1 analyzes the components of link collection LK of user A, which is the largest component of user E, and determines the user H (H (Recommended) is displayed. In this way, the values of the constituent elements are focused on the keyword “people”.
  • the main server on the Internet generated by XML or dynamic HTML (separate real-time or dynamic programs using various types of CGI, etc.) that can use the data separated and used. Synchronize the link collection (recommended site information of others, etc.) in another server in a distributed location that is not on (server 1) with the link collection (recommended site information, etc.) on server 1 and share the same data. Get and share in real time.
  • the link collection LK (a collection of link collections) of Server 1 is also updated synchronously.
  • the server 1 quantifies the personal value information into a dynamic Taste Code of a dynamic numerical value using a binary number or the like as in the processing of (5) above.
  • Taste Code can be encrypted, and communication with other groups (Sites A, B, C) becomes possible.
  • the server 1 uses this content to communicate with other site servers, exchanges data (TasteCode), and provides and charges user preference information.
  • TasteCode data
  • Each of the other companies' sites A, B, and C grasps the preference of the user E from the pre-existing TasteCode correspondence table.
  • the server 1 uses the Internet information displayed on the basis of the collection of individual content generation use frequency priority classification links created by the above processing, and voice calls or voice over IP (voice calls over IP).
  • This allows the Internet to be linked with a support call center.
  • connect to a specific server including phone calls) with a mobile phone, etc., and talk with a partner (character, etc.) that has been artificially created using voice recognition technology (including preference recognition using artificial intelligence).
  • voice recognition technology including preference recognition using artificial intelligence
  • While acquiring the speaker's preference attributes including interests, ages, hobbies, values, etc.
  • marketing including market research, certain purposes, and guidance to manufacturers. It can be extended to be linked with the services provided by the above processes (1) to (15).
  • the server 1 in the value information providing system SYS can execute the following processes.
  • composition ratio of the information provider can be determined, and the composition ratio of the theme unit TU can be quantified and symbolized (absolute taste code / value code) to form a group (specific identification). It is possible to determine the absolute position within a group or the whole), and to find similarities and similarities in values (see Figure 13).
  • a “recommendation” is dynamically issued to the agent generated in (1) above using the TasteCode (see Fig. 14).
  • the agent created in (1) above can dynamically talk to the user and all kinds of communication using the TaskCode (see Fig. 15).
  • the agent generated in (1) above is dynamically issued a recommended advertisement and any recommendation information (see Fig. 16).
  • the agent is a tool (robot) for automatically generating a theme i-nit TU composed of one or more data units DU.
  • This agent can obtain information from other Theme Unit TUs and can also provide information to other Theme Unit TUs.
  • Server 1 allows user E to add a data unit DU from the Internet (entire Web) or to automatically extract and install a data unit DU from another theme unit TU.
  • the theme unit TU displayed on the terminal device used by the user E is composed of a plurality of data units DU. These data units DU are automatically sorted by the server 1 in order of use frequency.
  • the server 1 converts the character string of each data unit DU in the theme unit TU into a standard symbol such as English, creates a frequency distribution and the like, and ranks and weights.
  • Server 1 uses these strings. Is converted to English as a standard symbol, "tra V e 1, gour met, domestic travel, hot—spring, gourmet” is obtained.
  • the server 1 converts the data into English. , "Hokkaido, family, hot-spring, seasons, not-spring, gour me tspot”.
  • the data unit DU is
  • the server 1 uses the arrangement and weight of the character strings as a value code (TasteCode).
  • Sano 1 can determine the position of the theme unit TU of the user E based on the relative TasteCode (TC) obtained in this manner.
  • Server 1 calculates the relative Taste Code of user E to the relative Taste Code of the group of other users F, G, H, I, and J obtained in the same manner, and the relative Taste Code of user E.
  • the following is a specific example.
  • Each data unit DU that makes up the theme unit TU has flags F1ag C, B, C, and A, and it can be understood from whom (from which theme unit TU) server 1 has extracted and introduced.
  • the second stage DU2 (weight 0.5) is extracted and introduced from B's TU (F 1 ag B), B (0.5)
  • the third stage DU3 (weight 0.33) is extracted and introduced from C's TU (Fl ag C), C (0.33)
  • DU4 in the fourth row (weight 0.25) is extracted and introduced from A's TU (Fl ag A), A (0.25)
  • the next data unit DU of the user I's theme unit TU is displayed. Since the theme unit TUs are automatically arranged in the order in which they are frequently used, the top units in the ranking other than the data unit DUs already extracted and introduced to the user unit TU of user E are displayed in order.
  • the character string of the data unit DU extracted and introduced into the theme unit TU is analyzed in the same manner as the relative taste code (Fig. 11), and the data unit of user I's theme unit TU, which has the closest value, is analyzed.
  • G Display in order from DU (except for those already installed).
  • the next step is to display the data unit DU of the theme unit TU closest to the relative taste code of the theme unit TU in the above manner. ( Figure 11).
  • the process proceeds to the theme unit TU of the user G or F, and the next theme unit TU is similarly used sequentially from the one with the closest value.
  • the left one uses relative taste code (Fig. 11) for narrowing down and estimating values.
  • the other right one uses absolute taste code (Fig. 12).
  • the aim is to infer a sense of values.
  • the extended guess of the values is performed in the same manner as the refinement guess of the values, but using the absolute Taste Code instead of the relative Taste Code.
  • a conversation pattern is also a theme unit TU, and a specific pattern (age group, dialect, wording, habit, etc.) is a specific theme.
  • this data unit DU but also extracted from this relative unit of the theme unit TU, fitted into the conversation pattern, synthesized and displayed.
  • the character string input to the user input window is analyzed in the same manner, inserted into the character string by the above method, synthesized, and displayed in the conversation display window.
  • server 1 analyzes the character string in the second row in the same manner as the relative taste code (Fig. 11), and In addition, an advertisement pattern prepared in advance is displayed.
  • the content strings of user actions (such as clicks) are analyzed, and several advertisements and conversation advertisement patterns (advertisement types of theme unit TU), which have been analyzed using the above method, have a large weight (importance). (High) is inserted into the character string, and the message is combined and displayed in the advertisement display window.
  • a relative taste code of the theme unit TU is extracted, embedded in an advertisement and a conversation advertisement pattern, combined and displayed.
  • a character string input to the user input window is analyzed in the same manner, inserted into the character string by the above-described method, combined, and displayed in the advertisement display window.
  • Advertisement content is also the theme unit TU, and comments that indicate the content are prepared in advance as data units DU.
  • the method of continuing and connecting the theme unit TU of the advertisement and conversation advertisement pattern uses the same method as the procedure for issuing the recommendation shown in Fig. 13.
  • Server 1 can use an agent (for example, a character or a pet) to automatically generate value codes (Taste Codes) that analyze the user's preferences and digitize and encode them. Enables communication with other users who have the closest preference to the user.
  • an agent for example, a character or a pet
  • Value codes Tin Codes
  • a community can be formed around users I, A, etc., which means that a community of multiple preference groups around the user himself can be formed. Furthermore, since there is a TassteCode of the depth of the community, it can be specified as a numerical value.
  • the server 1 in the value information providing system SYS can execute the following processes.
  • a unit that links the action of seeing a certain page on the Internet with the value of the user and adds various attribute flags to the summary (title, comment, URL) of this page.
  • the user's interests and preferences are entered using the DU) as the minimum unit (see Fig. 24).
  • a 1-to-1 (one-to-one correspondence) bookmark collection (link collection) is automatically generated for each user and tailored to the user.
  • the optimal arrangement is realized, and the user's sense of values is grasped and inferred for the dynamic movement of the data unit DU in this arrangement in real time (see Figs. 22 to 26).
  • the data unit DU is processed with certain values (see Fig. 24).
  • the data code of the DU obtained from the action of clicking (selecting) the data unit DU, changes in the contents of the DC and TU, TC, and the contents of the DU It captures changes in the position within the TU (frequently used DUs, etc.) and infers values (see Figures 22 and 23).
  • FIG. 17 shows an example of a display screen on the terminal device used by the user.
  • This screen consists of a reduced guess window, an enlarged guess window, an outline of the automatic bookmark collection, a search window, and a large frame.
  • the Taste Unit TU (a collection of value units and the smallest part data unit DU) is displayed.
  • this taste TU is a unit of data DU DU collection
  • this data unit DU is composed of title, comment, URL, and various F1ag, minimum unit)
  • T aS t e C o d e (general term for each T a s t e C o d e)
  • RTC Relative Taste Code (Absolute Value Code)
  • ATC Absolute Taste Code (Absolute Value Code)
  • the "people" factor is stronger than RTC
  • AIC ainascucode, where ainasu is the site name provided by server 1.
  • AIS AI site or partner site for art sourcing such as ASP (major classification)
  • STC TC of AIS
  • Character classification (This is called the middle classification. Characters such as pets and people are used to make this classification easier to understand.)
  • UBM U se r Book Ma r k (D U created by user)
  • AIC of user Ux STC + CTC + TC + PC.
  • the routing server has all character information such as AI site and ASP site including ainusu main server, index of TC and DC, and ID, Pass wor d information.
  • Category 1 has a relationship of AIS classification ⁇ character classification.
  • Fig. 18 shows a list (menu) screen of AI sites and its explanatory diagram.
  • the "AI site” menu is displayed only for the first time of user access.
  • the first "family” of the SDU is fixed at the first.
  • SDUs are ranked in order of frequency of use, except for “family,” and their averages change automatically.
  • Figure 19 shows the screen of the character “Taste Unit”, the character attribute, the data code DC, and the CDC.
  • DUx has the UX character code at that time, and the character attributes of CDU1, CDU2, CDU3, CDU4, and CDU5 (up to 5) added. Is done.
  • UBM DUx
  • the DUx has the character attributes CDU l, CDU 2 and CDU 3
  • UX will create its own character. Even if the attribute is changed, the UBM character attribute remains unchanged from the time of creation. This is because the interests and preferences of the creator at the time of creation are reflected.
  • the character attribute registration becomes 100 or more, and the character attribute in the DUx (UBM) created by this Ux is not in the character attribute registration.
  • this DUx is not reflected in character attribute search and sort. This is reflected only when “General” is selected.
  • search will be performed from all TUs (excluding CTU and STU).
  • the characters of "family” are, kindergarten (0-6 years old), elementary school student (7-12 years old), junior high school student (13-15 years old), high school student (16-18 years old), specialty and junior college. College students (over 19), OL / members (over 19), businessmen Z (over 19), moms, dads, grandpas and grandpas (over 60) There are 1 1 types.
  • Data Unit DU (DU 2) described the title “Tour of Hot Springs in Hokkaido Family” and the comment “Commentary on the recommended hot springs of the four seasons together with eating places”, “Hokkaido, family, hot — spring, seasons, hot - spring, it is gourmetspotj force s extraction.
  • the weight is set assuming that the title weight is 1 and the comment weight is 0.5. Then
  • each code such as CDC is coded in almost the same way.
  • each code has an absolute value Udo and a relative value code, and by using these codes comprehensively and weighting them (use priority, specific gravity, etc.), Ux's interests and preferences It is possible to capture values dynamically with an optimal and simple system.
  • the character name "gourmet, travel, domestic travel” Create a new character, CDU.
  • An illustration such as a person or an animal can be used for this, and the character such as the animal appears as a guide for the user who has selected the attribute.
  • Figure 20 is an explanatory diagram of the user Ux's automatic link collection screen and how to give recommendations by estimating reduction and enlargement.
  • the DU can display 100 DUs from DU 1 to DU 100 in total, "Mama” characters are displayed from DU 1 to DU 100. If the mom character is less than 1 00, a character that you selected in the AIS (in this case, "family", in the case of ainas u body whole and meaning) to display the total strike rankings to continuation of the mom character.
  • Figure 20 shows the top screen of Ux when “Mama” is pressed (specified) on the character screen, and when the “Search” button is specified, the DU in the TU of Ux (the DU pressed by Ux and the bookmark) Search for.
  • the right window of the magnified guess will display the next most established DU of the TU to which the mom's diary (DU) belongs.
  • the TC that has the highest value of the TC TU with the closest value to the TC will be displayed.
  • the next button will display the next highest DU.
  • use DUs in the TU's best ranking order in the case of an extended guess.
  • DUs are displayed in the order of the best ranking among the TUs with the closest values when arranged (sorted) around the TC of the TU to which this DU 1 belongs. This means that by displaying the DUs that the owner of DU 1 uses most often, the priority of display depends on the interests and preferences of this owner, and rather than narrowing down DU 1 to the center, DU 1 The key is to make extensive estimates based on the tastes of the original owner of DU 1.
  • Figure 21 shows a schematic diagram of the actual homepage display from the smallest unit, Dataunit DU.
  • Fig. 22 and Fig. 23 show the method of estimating the user's interest and preference by estimating reduction and expansion.
  • the left window displays the DU with the closest value when arranged around the DC of this DU 1.
  • DUs with similar values are displayed in order.
  • the right window displays the DUx of the TU's best ranking to which "Mama's Diary" (DU1) originally belongs.
  • the “recommended right window” DU is displayed in sequence. This means that the owner of DU 1 will display the most frequently used DUs, and the priority of the display will depend on the owner's interests and preferences. Instead of narrowing down DU 1 to the center, key DU 1 He speculated to expand along with the original owner's preference of DU 1.
  • Figure 24 solves the absolute value code ATC in the extended speculation of the recommendation. It is a figure for explaining.
  • the first stage DU 1 (weight 1) is extracted and introduced from C's TU (F l a g
  • the second stage DU2 (weight 0.5) is extracted and introduced from B's TU (F 1 ag B), B (0.5)
  • DU3 in the third row is extracted from C's TU (weight 0.33) and introduced (Fl ag C), C (0.33)
  • DU4 in the fourth row (weight 0.25) is extracted and introduced from TU of A (Fl ag A), and A (0.25)
  • the DUs are displayed in the order of the best ranking in the TU of the AT C that has the closest value when the AT Cs of this UX are arranged according to certain rules. This means that the ATC displays the DUs that are used most often by the owners of the closest TUs, and the display priority is JP01 / 05530
  • Figures 25 and 26 are illustrations of how to collect and estimate data for reduction and expansion estimation. .
  • a relative sort (sort) array is created by a certain rule centered on DCX.
  • n is the unique value of DC selected by sorting
  • DC (data code) is the value code of the smallest unit handled by this system.
  • DC x is displayed in the left window of the reduced guess with the center closer to DC x,
  • DC n (DC x + 1 1)
  • n is the eigenvalue of the TC selected by sorting
  • TU Taste Unit
  • TC X + 3 the unit of the smallest unit DU that is handled by this system
  • TC X-3 the value code
  • T C x This is displayed in the right window of the guess at the center, which is closer to TTC X than the TTC X.
  • Figure 27 shows a configuration example of the character screen.
  • Figure 28 is an explanatory diagram when there are multiple character attributes.
  • the DU 1—DU 3 is the DU pressed by Ux
  • the DU 4—DU 5 2 is the ma character (50% of 97)
  • Fig. 29 is a list of AI sites operated during use (menu) screen and explanatory diagram
  • Fig. 30 is a list of AI sites (menu) that is operated by a newspaper company
  • Fig. 31 is a newspaper company.
  • Fig. 32 shows a schematic diagram of the actual homepage display from the smallest unit of data unit DU at the AI site
  • Fig. 33 shows the character screen and taste when introduced to a newspaper company. It is a relational diagram of a to-unit (theme unit) TU.
  • Fig. 34 is an explanatory diagram of how to estimate the user's interests and preferences by inferring reduction and enlargement when introduced to a newspaper company, and shows the top screen of UX when "political economy" is pressed on the character screen. ing.
  • the left window displays the DUs with the closest values when arranged around the DC of this DU 1.
  • DUs with similar values are displayed next to each other. This means that for one interest shown by Ux, that interest is narrowed down (estimated by reduction).
  • the right window normally displays DUs in the order of the highest ranking in ⁇ , which has the closest value when arranged centering on the TC of the (original) TU to which this DU 1 belongs.
  • the data source TU of this AIS is only the theme unit, if the TUs of the UX are arranged around the TU of the UX instead of the TU to which this DU 1 belongs, DUs are displayed in the best ranking order.
  • the TU was originally only a theme cut, it became a similar data block.
  • Figure 35 shows a method of confirming conversational interest and reflecting it on the taste code TC. Since the DC (and TC) of the TU of the UX is formed from a set of nouns, this noun is inserted into a specific conversation pattern to confirm U's interest, and the answer is used as the weight of the original noun as DC To reflect.
  • the mood element (Sympathy Code SC) is an auxiliary code that expresses the person's mood and interest at that time, and is used to more accurately grasp the person's interests and preferences. It is a code that disappears with.
  • this code represents the current mood of the Ux when the DU was created.
  • the numbers are the respective weights, and the SC part disappears with time.
  • each process in each of the above-described embodiments can be performed by selecting and combining a plurality or all of the processes.
  • AI human-based artificial intelligence

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Information Transfer Between Computers (AREA)

Abstract

Un procédé de fourniture d'informations relatives au sens des valeurs comprend les étapes consistant à déterminer la priorité selon la fréquence d'utilisation, à classifier des pages Web, à atteindre un ensemble de liens originaux d'un fournisseur de données commentées, et à créer un ensemble de liens des contenus d'affichage d'utilisateurs individuels; et les étapes consistant à analyser les contenus de l'ensemble de liens par le procédé de déduction d'une intelligence de groupe, et à afficher l'ensemble de liens correspondant aux goûts des utilisateurs individuels sur un dispositif terminal utilisé par l'utilisateur.
PCT/JP2001/005530 2000-06-27 2001-06-27 Systeme fournissant des informations relatives au sens des valeurs WO2002001411A1 (fr)

Priority Applications (1)

Application Number Priority Date Filing Date Title
AU2001266343A AU2001266343A1 (en) 2000-06-27 2001-06-27 Sense-of-value information providing system

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
JP2000-234981 2000-06-27
JP2000234981 2000-06-27
JP2000311276 2000-09-04
JP2000-311276 2000-09-04
JP2000326204 2000-09-19
JP2000-326204 2000-09-19

Publications (1)

Publication Number Publication Date
WO2002001411A1 true WO2002001411A1 (fr) 2002-01-03

Family

ID=27344248

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2001/005530 WO2002001411A1 (fr) 2000-06-27 2001-06-27 Systeme fournissant des informations relatives au sens des valeurs

Country Status (2)

Country Link
AU (1) AU2001266343A1 (fr)
WO (1) WO2002001411A1 (fr)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003345827A (ja) * 2002-05-23 2003-12-05 Hewlett Packard Japan Ltd ポータルサイト最適化システム並びに関連する装置、方法及びデータ構造
JP2006244028A (ja) * 2005-03-02 2006-09-14 Nippon Hoso Kyokai <Nhk> 情報提示装置及び情報提示プログラム
JP2014021608A (ja) * 2012-07-13 2014-02-03 Sony Computer Entertainment Inc 処理装置
WO2016111065A1 (fr) * 2015-01-09 2016-07-14 ソニー株式会社 Système de traitement d'informations, dispositif de traitement d'informations, procédé de commande, et programme

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH113356A (ja) * 1997-06-13 1999-01-06 Nippon Telegr & Teleph Corp <Ntt> 情報共助方法及びシステム及び情報共助プログラムを格納した記憶媒体
JP2000048046A (ja) * 1998-05-29 2000-02-18 Sony Corp 情報処理装置および方法、並びに記録媒体

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH113356A (ja) * 1997-06-13 1999-01-06 Nippon Telegr & Teleph Corp <Ntt> 情報共助方法及びシステム及び情報共助プログラムを格納した記憶媒体
JP2000048046A (ja) * 1998-05-29 2000-02-18 Sony Corp 情報処理装置および方法、並びに記録媒体

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
JAMES RUCKER AND MARCOS J. POLANCO: "Siteseer: Personalized navigation for the web", COMMUNICATIONS OF THE ACM, vol. 40, no. 3, March 1997 (1997-03-01), pages 73 - 75, XP002944741 *
MIKIHIKO MORI, SEIJI YAMADA: "Bookmark agent: Bookmark no kyouyuu ni yoru joho kensaku no shien", DENSHI JOHO TSUUSHIN GAKKAI RONBUNSHI D-I, vol. J83-D-I, no. 5, 25 May 2000 (2000-05-25), pages 487 - 494, XP002944742 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2003345827A (ja) * 2002-05-23 2003-12-05 Hewlett Packard Japan Ltd ポータルサイト最適化システム並びに関連する装置、方法及びデータ構造
JP2006244028A (ja) * 2005-03-02 2006-09-14 Nippon Hoso Kyokai <Nhk> 情報提示装置及び情報提示プログラム
JP2014021608A (ja) * 2012-07-13 2014-02-03 Sony Computer Entertainment Inc 処理装置
US9805096B2 (en) 2012-07-13 2017-10-31 Sony Interactive Entertainment Inc. Processing apparatus
WO2016111065A1 (fr) * 2015-01-09 2016-07-14 ソニー株式会社 Système de traitement d'informations, dispositif de traitement d'informations, procédé de commande, et programme
CN107111648A (zh) * 2015-01-09 2017-08-29 索尼公司 信息处理系统、信息处理装置、控制方法和程序
JPWO2016111065A1 (ja) * 2015-01-09 2017-10-12 ソニー株式会社 情報処理システム、情報処理装置、制御方法、およびプログラム

Also Published As

Publication number Publication date
AU2001266343A1 (en) 2002-01-08

Similar Documents

Publication Publication Date Title
Rahman et al. A personalized group-based recommendation approach for Web search in E-learning
CN104254852B (zh) 用于混合信息查询的方法和系统
CN102163198B (zh) 提供新词或热词的方法及系统
CN101329674A (zh) 一种提供个性化搜索的系统和方法
US20050240580A1 (en) Personalization of placed content ordering in search results
JP5848199B2 (ja) 影響力予測装置、影響力予測方法、及びプログラム
Sisodia et al. Web usage pattern analysis through web logs: A review
KR20160055930A (ko) 연속적인 소셜 커뮤니케이션에 사용되는 콘텐츠를 능동적으로 구성하기 위한 시스템 및 방법
CN101111837A (zh) 查询自动分类的搜索处理
CN108460082A (zh) 一种推荐方法及装置，电子设备
CN101283353A (zh) 通过分析标签找到相关文档的系统和方法
CN105160545A (zh) 投放信息样式确定方法及装置
CN115526590B (zh) 一种结合专家知识和算法的高效人岗匹配与复推方法
JP5910316B2 (ja) 情報処理装置、情報処理方法、および、プログラム
KR20200133976A (ko) 콘텐츠 큐레이션 방법 및 장치
CN119357470A (zh) 基于人工智能的科普教育资源推荐系统
US20170109411A1 (en) Assisted creation of a search query
CN111339429B (zh) 一种资讯推荐方法
CN103262079A (zh) 检索装置、检索方法、检索程序、及存储该程序的计算机可读取记录介质
JP4308683B2 (ja) ユーザ活動履歴可視化・分析方法、ユーザ活動履歴可視化・分析装置、および、プログラム
WO2002001411A1 (fr) Systeme fournissant des informations relatives au sens des valeurs
JP6696270B2 (ja) 情報提供サーバ装置、プログラム及び情報提供方法
JP2006338086A (ja) 話題規模管理装置
CN113849518A (zh) 辅助搜索查询
KR101132431B1 (ko) 관심 정보 제공 시스템 및 방법

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AE AG AL AM AT AU AZ BA BB BG BR BY BZ CA CH CN CR CU CZ DE DK DM DZ EE ES FI GB GD GE GH GM HR HU ID IL IN IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MA MD MG MK MN MW MX MZ NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT TZ UA UG US UZ VN YU ZA ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): GH GM KE LS MW MZ SD SL SZ TZ UG ZW AM AZ BY KG KZ MD RU TJ TM AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE TR BF BJ CF CG CI CM GA GN GW ML MR NE SN TD TG

121 Ep: the epo has been informed by wipo that ep was designated in this application
DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
ENP Entry into the national phase

Ref country code: JP

Ref document number: 2002 506476

Kind code of ref document: A

Format of ref document f/p: F

REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase