Japanese-Chinese Information Retrieval with An Iterative Weight-tuning Scheme

Abstract

This paper describes our Japanese-Chinese cross language information retrieval system. We adopt “query-translation” approach and employ both a conventional Japanese- Chinese bilingual dictionary and Wikipedia to translate query terms. We propose that Wikipedia can be regarded as a good dictionary for named entity translation. According to the nature of Japanese writing system, we propose that query terms should be processed differently based on their written forms. We use an iterative method for weight-tuning and term disambiguation, which is based on the PageRank algorithm. When evaluating on the NTCIR-5 test set, our system acheives as high as 0.2217 and 0.2276 in relax MAP (Mean Average Precision) measurement of T-runs and D-runs.