A heuristics-based keyword and phrase ranking in the text corpus for question answering systems

This work presents an effective keyword extraction and ranking approach based on multiple text-based features from the text corpus for Question Answering systems by keeping the context and discourse of the text in view. Here, the extraction and ranking process is automatic. The approach uses data semantics and relationships to mark the keyword boundaries and intelligently annotates texts, uses various relationships among entities in the text and proposes various heuristic rules to calculate the ranking score of the keywords and phrases. The proposed approach can be integrated into information retrieval based system as well as in Question Answering systems. The current work is compared against other standard baseline methods to show its competence. The results achieved and the comparison with the various existing methods show the effectiveness and usefulness of the proposed model.