Information retrieval is a complex process that involves understanding the user’s requirements to provide appropriate and relevant results. This involves combined working of several techniques such as contextual analysis, correlation analysis, sentiment analysis and a good understanding of the user’s profile. This paper presents an effective relevance feedback based information retrieval model that aids in effective retrieval and organization of results such that information relevant to the users are given high priority. The user’s profile is constructed and reinforced with their queries and selection responses. This is iteratively performed such that the user’s profile gets strengthened with better and more appropriate rules. Result organization is performed based on the significance levels, sentiment and user’s preferences. Experiments on STS Gold Sentiment Corpus indicate effective predictions when compared with recent models.
Digital Object Identifier (DOI)
Subitha, S. and Sujatha, S.
"User Profile based Information Retrieval Incorporated with Reinforcement Learning,"
Applied Mathematics & Information Sciences: Vol. 11:
5, Article 26.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol11/iss5/26