Pattern Recognition and Stylometry Analysis of Pathittrupathu in Tamil Literature

This research paper attempts to identify the structure and clustering of Sangam Literature Pathittrpathu in Tamil using Data Mining Techniques. Pathittrupathu means, it is split into ten tens poems, which is a poetic work and one of the eight anthologies available in ancientTamil literature. This work originally contains ten divisions of ten poems each dedicated tovarious Chera kings rule of Tamil, except the poetic works of first and the last ten poems have been lost in before sangam period of Tamil. The remain eight poetic works is considered as a database in this research paper. Twenty parameters were used to identify the structure and pattern of Sangam literature of Tamil. Initially, usage of words is identifiedusing descriptive statistics and pattern of poetic structure is traced by using hierarchical data mining techniques. The result shows that all the eight Pathittrupathu split into two cluster and their writing styles were unique. The clustering methods achieved same clusters and named as natural clusters.