Browse Prior Art Database

Method of Virtual Review

IP.com Disclosure Number: IPCOM000033685D
Original Publication Date: 2004-Dec-23
Included in the Prior Art Database: 2004-Dec-23
Document File: 2 page(s) / 286K

Publishing Venue

IBM

Abstract

This invention solves the problem of collecting inputs from users about certain subjects (e.g person, products, etc.) from an extremely large number of people in significantly limited time

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 52% of the total text.

Page 1 of 2

Method of Virtual Review

The internet and similar sources of knowledge contain a wealth of information related to people's opinions about other people, such as politicians, or products, such as cars. It would be great if we can use all the information on the web to decide what people think about a certain candidate for the presidency of the United States, or about a certain car design recently announced. In general, the real purpose of polls and reviews is to get a good indicator of people's opinions about a certain matter. The relevance of polls depends greatly on the number of people questioned and opinions collected. The more the better. The ideal poll would be 100% complete and 100% accurate. But most polling systems collect information from a very small percentage of entire population.

This invention solves the problem of collecting inputs from users about certain subjects (e.g person, products, etc.) from an extremely large number of people in significantly limited time. This invention solves the problem of collecting inputs/opinion's about a certain subject (e.g. person or product) from an extremely large number of people in significantly limited time. This is done through an information analysis system called the virtual poll. The invention, Virtual Poll system, is applied to knowledge produced by people and published on the internet and similar mediums.

INTERNET

INTERNET

Crawler

Crawler

documents

documents

1

1

2

2

Target subject

List of positive words

List of negative words

Target subject

List of positive words

List of negative words

Find mentions of target subject

Find mentions of target subject

3

3

4

4

Scan neighborhood of mentions for

Scan neighborhood of mentions for

Score the target subject

Score the target subject

5

5

This invention is divided in 5 steps:

Step1: Define glossaries of positive and negative terms clustered/grouped by application domains (e.g., medical domain, chemical domain, IT domain)

Step2: The Internet is crawled regularly by a traditional crawler.

Step3: The content collected is presented to the next processing module a document at a time.

Step4: For each doc...