Abdelniser A. Mooman, Otman Basir, and Abdunnaser Younes
Intelligent Learning, Information Retrieval, Ontology, Tagging
Current search engines and information retrieval (IR) systems enable users to access the vast amount of information available on the Internet. However, owing to the size and dynamic nature of the information resources, the end user must sift through a large amount of retrieved information in order to find the desired information. To alleviate this difficulty, we present a novel approach for mapping users to the relevant information, by constructing specialized domains in which the users have demonstrated interest.We use reinforcement learning and social tagging in the approach to assist end users to select the concept domains most relevant to their needs. Our approach involves enriching the user's query with related linguistic ontologies and statistical semantic-related concept terms. We employ the Natural Language Process (NLP) techniques to enrich the user's query with semantic, lexical, synonymous terms, and probabilistic topic models, such as Latent Dirichet Allocation (LDA), to extract highly ranked topics from a query's retrieved information.
Important Links:
Go Back
