Analysis of N-gram model on Telugu document classification

Document classification is one of the recent areas of research evolved as a result of exponential growth in the quantum electronic form of documents. Various document representation methods based on linguistic knowledge are revisited in literature. Adaptability of N-gram models on various languages is the recent trend. In this paper an attempt is made to analyze character N-gram model on Telugu documents. Tokenization of syllables and the associated complexity of Telugu script is described. A combination of Bayes probabilistic classifier and character N-gram model is discussed in this paper. The performance of the proposed classifier is evaluated in terms of overall accuracy.