Browse Prior Art Database

A Method and Apparatus for Event Trend Tracking

IP.com Disclosure Number: IPCOM000190598D
Original Publication Date: 2009-Dec-08
Included in the Prior Art Database: 2009-Dec-08
Document File: 3 page(s) / 250K

Publishing Venue

IBM

Abstract

When we find information from mails, webs, search engines, blogs, and BBSs, what we get is separated papers or very long lists of search results. It is far away from satisfying us for the needs of information. Now a day, we face the information explosion, and it is becoming harder to catch the detail tread of a topic. Strange things happen every day; Google has “Google Treads” for knowing the past and current of a query. We know from it that what is the history of the query talks about, and the hotness in its history.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 55% of the total text.

Page 1 of 3

A Method and Apparatus for Event Trend Tracking

We begin with an example coming from Sina.com:

Sina Topic

Summary of the Invention:

Our invention proposes an automatic method and system for tracking major the threads and branches when the whole thing grows and develops. The main idea is to apply extracting subject-predicate relationship and time spectrum to generate tree like threads to summarize all the branches of a topic.

The benefits of our method are as follows:

¨C More details provided within the narrow returned page
¨C Know the whole story by a fast glance;
¨C High lighted threads for quick result browsing;
¨C User could know the whole story of his query pointing to, which is very valuable to meet the user's search requirement. And he does not need to think how to tune the queries to get more aspects of his requirements.

1

[This page contains 1 picture or other non-text object]

Page 2 of 3

Description of the Invention:

The whole process of tracking a topic represented by a query and all its branches is as the following diagram:

Figure

1. The whole process of our invention

Followings are the key components and their functions in the system:

Subject-predicate-object triple extractor: An algorithm that can automatically all the subject-predicate-object triples of an input content.

Relation Annotator: An algorithm that detects the relationships between these subject-predicate-object triples. Relation detection can also be based on clustering schemas and models.

Branch propagation and tree generating and merging: This processor will detect which triple is bigger than others, who generate others, and recursively generate new branches.

¨C 1) generate: As a simple example: the triple B occurs after the day triple A occurred, and the document list containi...