Framework for Urdu News Headlines Classification

Abstract

Automatic text classification has great significance in the field of text mining and plays a pivotal role in areas such as spam filtering, news classification, noise reduction etc. It is evident from the literature that there is ample of research conducted for classifying text documents e.g. English news classification, Persian text classification etc. but there is no copious amount of work related to short Urdu text or Urdu news headlines classification. Therefore, after examining various existing news classification methodologies we propose an SVM based framework in this paper for classification of Urdu news headlines. This approach classifies Urdu news based on headlines in their respective pre-defined categories by utilizing their feature vector’s maximum indexes. This proposed system is compared with existing state-of-the art techniques.

Keywords

• Urdu News Classification
• Feature Vector
• SVM