Comparison of Genre based Tamil Songs Classification using term Frequency and Inverse Document Frequency Kanchana Sundara1,*, Meenakshi K.2,**, Ganapathy Velappa3,*** 1Academic Administrator, School of Computing, SRM University, Kancheepuram-603203, Tamil Nadu, India 2Assistant Professor, School of Computing, SRM University, Kancheepuram-603203, Tamil Nadu, India 3Professor, School of Computing, SRM University, Kancheepuram-603203, Tamil Nadu *Corresponding Author E-mail: kanchana.j@ktr.srmuniv.ac.in
**meenakshi.k@ktr.srmuniv.ac.in
***ganapathy.v@ktr.srmuniv.ac.in
Online published on 17 July, 2017. Abstract Genre classification of Tamil songs is done based on mood, emotion etc. of the listener. The musical genre classification is based on three levels as base, mood and style. Our proposed methodology is comparison of genre based Tamil songs classification using Term Frequency and Inverse Document Frequency (tf-idf). So for this methodology has been applied for English and Korean songs only. In this paper, we are classifying the Tamil songs using the method based on tf-idf scores. The classifier establishes a relation between the features of the training samples and related categories. The frequency of word usage is identified by term frequency and inverse document frequency. Support Vector Machine (SVM) algorithm and Naïve Bayes algorithm (NB) are used in Weka classification tool for analysis. We have compared the experimental results of both the algorithms in musical genre classification. From the experimental values obtained using various parameters, we have shown that the Naïve Bayes algorithm has classified the genre marginally better than Support Vector Machine algorithm. We have considered 2000 Tamil songs for testing the genre classification. At present we have considered two genres for classification and the same classification can be extended to other genres in future. Top Keywords Tf-idf, genre, lyrics, Tamil songs, base, mood, style. Top |