Dictionary-based Thai CLIR: Experimental survey of Thai CLIR

Abstract

This paper describes our work, which participated in the Cross-Language Information Retrieval (CLIR) at the Cross-Language Evaluation Forum. Our objectives for this experiment have three folds. Firstly, the coverage of the Thai-bilingual dictionary was evaluated when translating queries. Secondly, whether the segmentation process has effected the CLIR. Lastly, this research investigates the query formations techniques. Since this is the first international experimental in CLIR, our approach used dictionary-based technique to translate Thai queries into English queries. Four runs are submitted to the CLEF: (a) single mapping translation with manual segmentation, (b) multiple mapping translation with manual segmentation, (c) single mapping translation with automatic segmentation and (d) Single mapping with query enhancing with the Thai thesaurus words. The retrieval effectiveness is worse than our expected. The simple dictionary mapping technique is unable to achieve the retrieval effectiveness, although the dictionary lookup gave very good high percentage of mapping word. The words from the dictionary lookup are not specific terms but each is mapped to a definition or meaning of that term. Furthermore, Thai stopword, stemmed word and word separation have effected in Thai CLIR.