Combination Methods for Improving the Reliability of Machine Translation Based Cross-Language Information Retrieval

摘要

Cross-Language Information Retrieval (CLIR) is an important topic in the increasingly multilingual environment of online information. Experiments using the standard CLEF 2001 bilingual task show that Machine Translation (MT) can provide effective search topic translation for CLIR, and that retrieval performance can be improved and made more reliable by applying a combination of pseudo-relevance feedback, corpus methods and data merging.