REAL-TIME CONTEXT BASED INFORMATION RETRIEVAL USING SEMANTIC AND FEATURES ON STREAMING CONTENT

N.Karthikeyan, R.Dhanapal

Abstract: Information retrieval has become one of the major necessities in the current information age. Appropriate information retrieval is still complex due to data heterogeneity and unconventional modes of data dissemination. This paper presents a context based information retrieval architecture to perform real time retrieval of data. This model operates on the semantic similarity of the data, rather than the content similarity. Hence this technique exhibits better and efficient retrieval levels providing high reliability. The retrieved documents are pre-processed and feature vectors are created from the tokens. Polarity matching is used to filter semantically correlated results and magnitude based result ranking is performed to perform further elimination of inappropriate results. Experiments conducted using the proposed model exhibits very high true retrieval rates, along with high precision and recall levels exhibiting the competence of the proposed work.

Keywords: Information Retrieval, Context Based Data Analysis, Semantic Correlation, Polarity Identification, Sentiment Analysis

DOI: https://doi.org/10.15623/ijret.2016.0512019