ITRANSED MARATHI LITERATURE RETRIEVAL USING SMS BASED NATURAL LANGUAGE QUERY

SMS based information retrieval is one of the challenging research driven application of information and communication technology. “A flexible natural language query in the form of an SMS could also be answered by applying information retrieval theory and practices”, is our research theme. In the initial phase, we have formulated a preparatory experimentation on SMS based Marathi query to retrieve Marathi literature documents. This paper presents the details about the ITRANS Marathi literature documents followed by the description of an appropriate Vector Space Model (VSM) to represent the documents and the queries. We designed a GUI and asked a group of hundred students to formulate an SMS query for 30 different questions in order to obtain user specific variations among typed in queries. A collection of sample Marathi documents in the ITRANS form and a set of acquired queries is processed using the system developed to retrieve relevant documents. The system uses the Cosine Similarity to rank the documents according to the relevance for each selected query. Experimental results of the system are presented in the paper. This is the first such type information retrieval model for the ITRANS Marathi documents.