Improving Identification of Latent User Goals through Search-Result Snippet Classification

Abstract

In this paper, we propose an enhanced approach to improving our previous method which employs syntactic structures (verb-object pairs) to identify latent user goals. Our new approach employs a supervised-learning method to learn hint verbs and considers URL information and title information to classify snippets into three coarse categories, which are resource-seeking, informational, and navigational. Also, we propose three different models to identify three different categories of specific latent user goals from the classified snippets.