Abstract | ||
---|---|---|
This paper investigates our experimental results for NTCIR-4 QAC2, the second attempt to evaluate the technology of Japanese question answering (QA). Our basic approach is a combination of information retrieval and named entity (NE) extraction based on pattern matching. The results show that the accuracy of NE extraction crucially affects the overall performance of our system. Additional experiments show the effects of refinements of answer extraction. We also analyze the QAC2 test collection to identify features relevant for measuring the difficulty of the questions in the collection. Based on the analysis, we make some proposals for the future QAC tasks, as regards to the measurement of difficulty of the test collection and definition of tasks. |
Year | Venue | Keywords |
---|---|---|
2004 | NTCIR | information retrieval,question answering qa,named entity extraction,pattern matching |
Field | DocType | Citations |
Question answering,Information retrieval,Computer science,Named entity,Pattern matching | Conference | 3 |
PageRank | References | Authors |
0.83 | 3 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Masako FUKUSHIGE | 1 | 3 | 0.83 |
Yoshio SATO | 2 | 3 | 0.83 |
Mitsuhiro SUZUKI | 3 | 6 | 1.28 |
Hiroyuki Suzuki | 4 | 3 | 0.83 |