JISE

[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24]

Journal of Information Science and Engineering, Vol. 26 No. 2, pp. 685-697

Japanese-Chinese Information Retrieval With an Iterative Weighting Scheme

CHU-CHENG LIN, YU-CHUN WANG⁺ AND RICHARD TZONG-HAN TSAI⁺⁺
Department of Computer Science and Information Engineering
⁺Department of Electrical Engineering
National Taiwan University
Taipei, 106 Taiwan
⁺⁺Department of Computer Science and Engineering
Yuan Ze University
Chungli, 320 Taiwan

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 weighttuning and term disambiguation, which is based on the PageRank algorithm. When evaluating on the NTCIR-5 test set, our system achieves as high as 0.2217 and 0.2276 in relax MAP (Mean Average Precision) measurement of T-runs and D-runs.

Keywords: Japanese-Chinese cross-language information retrieval, query disambiguation, iterative term weighting, query-translation, named entity translation

Retrieve PDF document (JISE_201002_22.pdf)