A New Clustering Algorithm for Comparable Entities from Web

Abstract

In internet comparison activity performed by users for decision making .It is very difficult what to compare and what are alternatives. The comparable entities can be used to help users make alternate decisions by comparing relevant mining entities. Several approaches exist to extract comparable entities from various web corpuses. Existing entity mining techniques focus on mining comparable pairs readily observed in the web corpus. a weakly-supervised bootstrapping method can be used to identify comparative questions, comparative patterns, and extract comparable entities. But our work focuses on predicting pairs that cannot be observed from it. For this we develop TricluQueue clustering approach for comparative question identification and comparable entities extraction. We aim to find clusters in which all entities within the same cluster are comparable to each other.