Context-based News Articles Retrieval using CLSM

Abstract

With the continuous growth of the electronic data and the expansion of World Wide Web, users are flooded with information for a single search query. Most commonly, the results that arrive for the search query ignore the context. Context has several dimensions such as users context, query/document context, spatial/temporal context. Though context is not a new idea in building intelligent systems, Contextual Information Retrieval is a biggest challenge in IR domain.This paper proposes news article retrieval using Convolutional Latent Semantic Model (CLSM). CLSM extracts the contextual features present in the query and finds the relevant documents and ranks them based on their relevance with the given query. CLSM was experimented with the clickthrough data of a commercial search engine and has been proven for its context-sensitive results and efficiency. In this paper, we discuss the feasibility of using CLSM for extracting news articles based on the context present in the query from a static news article repository.