Available Methods in Farsi-English Cross Language Information Retrieval Using Machine-readable, Bilingual Glossary

Abstract

In this paper the impact scope of Natural Language Processing (NLP) on translating search statements was determined by testing out research hypotheses. The NLP techniques employed for search statement processing included text parsing, linguistic forms identification, stopword removal, morphological analysis, and tokenization. Examination of the hypotheses indicated that using the method of translating the first equivalent term selected versus the method of selecting all equivalent terms, would contribute to increased efficiency of the review that while morphological analysis of the terms not translated by the glossary, would increase the retrieval precision cutoff, there would be no significant difference established by the lack of such analysis thereof that sentence translation as opposed to term by term translation, would increase the efficiency of Farsi-English proofreading. Other findings are also represented.

Keywords

• Cross Language Information Retrieval
• Natural Language Processing
• Machine Readable Bilingual Dictionary
• Information Retrieval Evaluation