Information routing

Details Information routing Information routing

1999-01-14

2001-07-24

Alam, Hosain T. (Department: 2172)

Data processing: database and file management or data structures

Database design

Data structure types

C707S793000, C707S793000, C709S203000, C709S206000, C709S228000, C713S176000, C379S220010, C379S265080

Reexamination Certificate

active

06266667

ABSTRACT:

This application claims priority under 35 U.S.C. §§119 and/or 365 to 9800076-3 filed in Sweden on Jan. 15, 1998; the entire content of which is hereby incorporated by reference.
TECHNICAL FIELD OF INVENTION
The present invention relates in general to a method for finding and retrieving relevant electronic information and in particular to a method for finding and retrieving relevant electronic information from the Internet.
DESCRIPTION OF RELATED ART
Today the Internet is growing in a tremendous way, and the amount of information is overwhelming for the ordinary user. A big problem for a user is to find relevant information. There is also a problem for content providers to reach out with the information they want to deliver.
In the recent years a new paradigm has emerged on the Internet, the Software Agent. The users of such agents are supposed to get help to find relevant information. The agents are used in areas called Information Retrieval and Information Filtering. Often they use techniques from machine learning.
The Problem Area
When a user wants to find interesting documents on the WWW (World Wide Web), an active search of the web is necessary to retrieve the needed information. Some tools to help the user find relevant documents exists, but most of these tools use query-based techniques. The query-based techniques has properties which give results which include many irrelevant documents and exclude many relevant documents.
The usual way to get around this problem is to use different retrieval and filter techniques that are able to adapt to the user automatically or manually.
Hereinafter, a consumer is somebody who wants information, a content provider is somebody who provides information and artificial evolution is a machine learning technique that is inspired by evolution biology and uses mutations and reproduction to create programs. A consumer or a content provider may also be termed a user.
Filtering and Retrieval
There are mainly two ways to filter and retrieve information, the cognitive and the social techniques.
The cognitive or content-based approach analyses the content of the information and compares it to a model of the user. A closer match makes it more likely it will suit the user.
The social approach determines what is regarded as relevant information solely based on what different users are recommending. If a first user finds a particular kind of documents relevant, which other users found relevant too, then the first user will probably find one particular document relevant if the other users found it relevant.
There exist approaches that try to combine the above described two techniques. The usual information retrieval techniques are based on query refinement. The system uses the user's feedback to refine the query and get a better retrieval of documents.
This is a form of user modeling and can be categorized as a cognitive technique.
Content-Based examples
In the content-based approach the information is collected by an agent according to a user-profile or user-model from the content provider and is presented for the user.
A cognitive filtering example is Beerud Seth's Personalized Information Filtering
A learning approach to personalized information filtering, Master Thesis, MIT.
It uses an artificial evolution approach to build a user-model (or user-profile). As genotype it uses the user-profiles, but it also lets the profiles learn during their lifetime to make the adaptation faster, which is called the “Baldwin effect”. Each profile search for documents and recommends them to the user. The user's response is used to change the fitness of the profile and to let it adapt during its lifetime. This approach, because of the evolutionary algorithm, tries to benefit from an effective parallel search.
An other example of the cognitive approach is The Info Agent D'Aloisi, D & Giannini, V
The info agent: an interface for supporting users in intelligent retrieval, Published on Internet.
The Info Agent consists of three different cooperating agents, the Interface Agent, the Internal Services Agent and the External Retrieval Agent. The Interface Agent builds a user-profile and it uses the profile to guide the two other agents' search for documents. The system is designed to be flexible to extensions and changes.
An example of information retrieval is Discover. Sheldon, M A & Duda, A & Weiss, R & Gifford, D K.
Discover: A resource discovery system based on content routing. Proceedings of the third international world wide web conference,
1995. This document describes an architecture for a single point of access to over 500 WAIS servers. It provides two key services, query refinement and query routing. The user uses query refinement to make the query more precise. When the query is precise the query is routed to a suitable WAIS server to retrieve the relevant documents. The WAIS servers have local databases of documents and they coo-operate through a content router that the user can access.
Social Examples
In the social approach the information is collected from the content provider and stored by a consumer in a central repository and than retrieved again for another consumer that might store it again.
A social filtering system for net news is presented in Maltz, A D.
Distributed information for collaborative filtering on usenet net news. Master thesis, MIT.
In this system, each user can read an article and vote for or against it. The votes are then sent to a vote server where the votes are grouped together and shared with other vote servers. The servers aggregate all the different readers' opinions into one collective opinion. This aggregated opinion is then used by news-readers to filter shown articles.
Firefly, http://www.firefly.net, is one of the best-known social filtering systems. The technique is called Feature-Guided Automated Collaborative Filtering. This filtering technique builds a profile of each user with their opinions for different documents. The documents are divided in different groups (classes of documents) and for each group the users are clustered in a nearest “neighbor” style. For each group of documents the users' opinions in the same cluster are compared. To find documents to recommend, firefly matches all users in a cluster (this is somewhat simplified to increase understandability). If two users have approximately the same opinions for most of the documents in a group, but they have not read all of them, then it is likely that the users would like the unread documents of the group too.
Combined Techniques
In combined techniques, the information flow is the same as for the social techniques.
Marko Balabanovic,
An adaptive web page recommendation service. Stanford digital library project working paper SIDL
-
WP
-1996-0041, describes an example which utilizes both filtering techniques. This architecture consists mainly of two kinds of agents, a selection agent (one for each user) and a collection agent. The collection agent collects information and deliver documents to a central repository, where from, the selection agents filter interesting documents according to a user-profile. Responses from the users are used to modify both the selection agent and the collection agent. If a document matched a user, the collection agent that found this document is “rewarded” and the agent is thus encouraged to specialize in a certain kind of documents. The user is also able to grade a sample of pages. The selection agent sends the pages that the user grades as very interesting to neighbor users with the same interests. The selection agent also handles the recommended pages in the same way as other pages from the repository. This means that if the pages do not match the user's profile they are not presented to the user, and this also then means that the gain from social filtering is very limited.
Collaborative Agents that Share Information
Yet another approach is where the information is collected from the content provider, which might be a consumer too, and then sent to another consumer, which might also be

Affiliated with

Also associated with

Rating

Information routing 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 routing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Information routing will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2539090