Information Retrieval Based on Extraction of Domain Specific Significant Keywords and Other Relevant Phrases from a Conceptual Semantic Network Structure

This paper presents a functional approach towards the problem domain of Information Retrieval System built upon a narration based search text. The presented system retrieves documents from the background collection by extracting the domain specific significant keywords and other relevant phrases from a given narrative search text. The narrative search text can be a description or scenario which poses a great difficulty in the problem domain to retrieve the relevant document sets with efficiency and accuracy from the background data repository. We have adopted a different approach where the significant keywords are extracted from the narration text to form a search query and alternative sets of queries are also formulated by expanding the search query from a Conceptual Semantic Network built for the purpose. Inclusion of the Conceptual Semantic Network and WordNet synonym sets for the search query expansion plays an important role in the retrieval mechanism. Experiments were carried out on the data sets from the Ad-hoc task of ‘Forum for Information Retrieval Evaluation, 2013’. The background data set contained a huge number of legal documents consisting of data over 3 GB and was divided into two domains such as ‘Consumer Law’ and ‘Hindu Marriage & Divorce Law’. For the search queries, a set of scenarios in the form of narrative text were provided. The system was required to perform an analysis of the search text and retrieve a set of top 1000 legal documents for each of the queries from the background collection which may be relevant to the situation described in the narration of the search text.