An Open Information Extraction For Question Answering System

An Open Information Extraction (OIE) is a process of extracting meaningful information in a structured format from a set of unstructured or semi-structured data. OIE is used for applications like question answering, bio-text mining, text comprehension, ontology learning and sentence similarity. The main focus of OIE is to extract maximum possible set of verb-based triples. However, it ignores extracting adverbial and adjective clauses that are necessary for efficient semantic applications like question answering and paraphrase identification. Our method deals with the use of OIE techniques for extracting clauses which includes the adverbial and the adjective clauses. We have evaluated our OIE approach to extract answers for the given questions. The answers for the questions are determined based on pattern matching. We have used New York Times data set to evaluated the performance of our question answering system using OIE approach. We have obtained precision, recall and F1-measure as 39.47%, 75% and 51.7% respectively.