Automatic extraction of attributes and entities for product differentiation

Abstract

This paper introduces a novel concept of web-spreadsheet, which extracts product information by crawling through related web pages and generates information like a spreadsheet where each row represents product information and each column represents product attributes. Using Decision Tree based classifier, the accuracy achieved for extracting entities and attribute-value pairs is 98.71% and 99.99% respectively. This strategy will benefit the product companies for performing successful product differentiation and save a lot of time of the customers for comparing different products of same product class.