ECIR'11 Proceedings of the 33rd European conference on Advances in information retrieval
| Hi-index | 0.00 |
Question Answering systems involve the extractionof answers to a question rather than retrieval of relevantdocuments. For Question Answering evaluation, it is necessarythat a human assessor decide the correctness of the answers,given that the same answer can be expressed in different ways.Therefore, the use of suitable test collection could help to identifywhere the systems are performing well, or where they are failing.Our data collection analysis suggests that there should be aproportion of text in which the reasoning or explanation thatconstitutes an answer to a “why” question is present in, orcapable of extracting from, the source text. We report on animplemented component for the extraction of candidate answersfrom source text. This component uses an approach thatcombines lexical overlapping and lexical semantic relatedness(lexico-syntactic approach) for ranking possible answers tocausal questions. On undifferentiated texts, we obtain an overallrecall of 34.13% indicating that simple match is adequate foranswering over 1/3 of “why” questions. We have analyzed thosequestion-answer pairs units where the answer is explicit,ambiguous and implicit, and shown that if we can separate thelast category, the rate of recall increases considerably.