Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
1999-11-01
2001-09-04
Black, Thomas (Department: 2171)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C709S222000, C709S218000
Reexamination Certificate
active
06286012
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to an information filtering apparatus and information filtering method for facilitating acquisition of necessary information from a storage using a magnetic, optical medium or the like and an information communication network.
2. Description of the Related Art
With progress of social infrastructure in recent information communications, it is desired greatly that an information filtering apparatus be achieved as a technology of supporting enlarged information communication networks and remarkable increases of an communication amount. As the background thereof, it is considered that the information amount which a person can access is greater than that the person can process. Therefore, the information which the person needs often lies buried in a great amount of information.
As a conventional technology relating to the information filtering apparatus, there is a key word logical expression used in, for example, patent search. That is, patents of several hundreds of thousands to several millions are subjected to filtering with the key word logical expression.
However, some problems have been pointed out in the conventional method using the key word logical expression. First, since it is necessary for a user to set a logical expression for a key word with high accuracy, the user needs to adequately recognize characteristics of a filed data group (for example, under what condition the key word of the data is determined) or a structure of a system (for example, whether or not the key word is the system with thesaurus system) to obtain fine retrieval. Secondly, the result obtained by the information filtering is only appraised as matching with the logical expression for the key word. Such a result only matches with the key word, and sometimes is different from the content that the user requires. Alternately, it is not easy for the user to acquire the information with high priority for the user in descending order of priority from among such many retrieval results.
Based on the forgoing, the inventors of the present invention developed an information filtering apparatus for enabling acquisition of filtered result information which is really necessary to the user with high accuracy. This information filtering apparatus is disclosed in Japanese Unexamined Patent Publication HEI9-288683.
The information filtering apparatus described in the above Japanese Patent Publication is explained with FIG.
1
. The information filtering apparatus is provided with information filtering unit
1
which subjects an input item to filtering, learning unit
2
which generates necessary information for the information filtering by learning, and interface unit
3
which indicates the filtered item on a display.
Information filtering unit
1
converts a plurality of key words contained in an item into a vector using a code dictionary stored in code dictionary storage section
4
. The information filtering unit
1
further calculates a score using the vector. For the score calculation, the positive metric and negative metric for a key word signal are respectively fetched from positive metric storage section
5
and negative metric storage section
6
. The necessity and reliability are calculated from the calculated score and a parameter stored in determination parameter storage section
7
.
When the necessity and reliability for the input item are thus obtained, the unread data items stored in unread data storage section
8
are rearranged in order of necessity. Interface unit
3
presents the items in descending order of necessity, while obtaining a user evaluation whether the presented item is necessary or necessary. Learning unit
2
modifies the contents of storage sections
4
to
7
for storing the matrixes for use in the score calculation using the user evaluation and the plurality of key words.
There is however a problem that the information filtering apparatus described above cannot divide items for each field thereof to present when user's interests include a plurality of fields.
SUMMARY OF THE INVENTION
An object of the present invention is to provide an information filtering apparatus and information filtering method capable of changing an ordering manner for the order of presenting items and dividing the items to be presented for each field thereof to present when predetermined items are acquired from a storage using a magnetic or optical medium and an information communication network.
In order to achieve the above object, the information filtering apparatus has a configuration having a metric dividing section which divides a matrix indicating that what item a user needed and did not need into a plurality of matrixes, and a classification filtering unit which calculates a plurality of classification necessities and a plurality of classification reliabilities using the plurality of matrixes divided in the metric dividing section.
According to such a configuration, it is possible to convert a plurality of key words which is a signal that is not defined as a distance into a vector expression that is defined as a distance using a metric in which the degree of necessity for the user is reflected, thereby to quantify the degree of necessity for the user. As a result, the user can obtain items in descending order of necessity, in particular, can obtain items in descending order of necessity for each of a plurality of classifications.
The first scope of the present invention is provided with a metric storage in which a determination result of necessary or unnecessary for a presented item is stored in the form of a metric signal calculated from a key word signal contained in the item, the metric dividing section which fetches the metric signal from the metric storage, and dividing the metric signal for each predicted interested field to generate a classification metric signal, a classification filtering section which rearranges the order of presenting input items which are input from a storage using a magnetic or optical medium, or an information communication network, using a divided metric signal, and a presentation section which presents the items for each interested field according to the rearranged order of presenting items. The first scope thereby has effects of classifying items using an input indicative of evaluation of necessary or unnecessary from the user to perform rearrangement of the items, and of presenting the items in descending order of necessity for the user.
The second scope of the present invention is to divide the metric signal, which is calculated from a key word signal contained in a presented item so as to indicate a determination result of necessary or unnecessary for the presented item, for each predicted interested field to generate a classification metric signal, and rearrange the order of presenting input items which are input from a storage using a magnetic or optical medium, or an information communication network, using a divided metric signal, and present the items for each interested field according to the rearranged order, thereby providing effects of classifying items and rearranging the items in order of necessity for the user with high accuracy.
The third scope of the present invention is a database restructuring method for restructuring a database using the information filtering method of the second scope, thereby providing an effect of enabling the user to acquire necessary items readily.
REFERENCES:
patent: 6076082 (2000-06-01), Kindo
patent: 6078916 (2000-06-01), Culliss
Kindo et al., Adaptive Personal Information Filtering System That Organizes Personal Profiles Automatically, IJCAI-97, v. 1, pp. 716-721, Aug. 23, 1997.*
Yoshida et al., “A Newspaper Reading Out System with an Adaptive Information Filtering Technology to Support Visually Impaired People,” IEEE, 1999, p. 38-41.*
Meki et la., “Competitive Model To Classify Unknown Data Into Hierarchical Clusters Through Unsupervised Learning,” IEEE, 1997, p. 815-818.*
English language abstract of JP 9-288683.
Kindo Toshiki
Watanabe Taisuke
Yoshida Hideyuki
Black Thomas
Greenblum & Bernstein P.L.C.
Matsushita Research Institute Tokyo, Inc.
Rones Charles L.
LandOfFree
Information filtering apparatus and information filtering... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Information filtering apparatus and information filtering..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information filtering apparatus and information filtering... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2498232