This has resulted in depriving us of the opportunity to use the language complexity to generate high-quality translations. Most of the existing approaches to machine translation (MT) circumvent the complexity with the assumption that morphological, syntactic and semantic analysis can be done independently and sequentially. Machine translation turns out to be an inherently complex process requiring serious attention to morphological, syntactic and semantic complexity within both the source and the target languages. The best MAP values for Bengali, Hindi and English queries for our experiment were 7.26, 4.77 and 36.49 respectively. The CLEF results suggested the need for a rich bilingual lexicon for CLIR involving Indian languages. Lucene framework was used for stemming, indexing, retrieval and scoring of the corpus documents. Other language-specific resources included a Bengali morphological analyzer, a Hindi stemmer and a set of 200 Hindi and 273 Bengali stop-words. We adopted Automatic Query Generation and Machine Translation approach for our experiment. Under this limited resources, we mostly depended on our phoneme-based transliterations to generate equivalent English query from Hindi and Bengali topics. But neither we had any effective Bengali-English bilingual lexicon nor any parallel corpora to build the statistical lexicon. For our experiment, we had access to a Hindi-English bilingual lexicon, 'Shabdanjali', consisting of approx. The cross-language task includes the retrieval of English documents in response to queries in two most widely spoken Indian languages, Hindi and Bengali. This paper describes our experiment on two cross-lingual and one monolingual English text retrievals at CLEF 1 in the ad-hoc track.

Bengali hindi letters