论文信息 - A Question Answering System based on Information Retrieval and Validation

A Question Answering System based on Information Retrieval and Validation

Our participation at ResPubliQA 2010 was based on applying an Information Retrieval (IR) engine of high performance and a validation step for removing incorrect answers. The IR engine received additional information from the analysis of questions, what produces a slight improvement in results. However, the validation module discarded sometimes too much correct answers, contributing to reduce the overall performance. These errors were due to the application of too strict constraints. Therefore, future work must be focused on reducing the amount of false negatives returned by the validation module. On the other hand, we observed that IR ranking offers important information for selecting the final answer, but better results could be obtained if additional sources of information were also considered.

Anselmo Peñas | Álvaro Rodrigo | Joaquín Pérez-Iglesias | Guillermo Garrido | Lourdes Araujo

[1] Lourdes Araujo,et al. Approaching Question Answering by means of Paragraph Validation , 2009, CLEF.

[2] Lourdes Araujo,et al. Information Retrieval Baselines for the ResPubliQA Task , 2009, CLEF.

[3] M. Felisa Verdejo,et al. The Effect of Entity Recognition in Answer Validation , 2006, CLEF.

[4] Vladimir I. Levenshtein,et al. Binary codes capable of correcting deletions, insertions, and reversals , 1965 .

[5] I Levenshtein Vladimir. BINARY CODES CAPABLE OF CORRECTING DELETIONS, INSERTIONS, AND REVERSALS , 1966 .

[6] Stephen E. Robertson,et al. Some simple effective approximations to the 2-Poisson model for probabilistic weighted retrieval , 1994, SIGIR '94.