Natural Language Processing an Efficient Tamil Text Compaction System

Tamil is slowly becoming the online language and mobile text messaging languages for many Tamils around the world. Social networks and mobile platforms now extensively support Unicode and applications for keying Tamil text. The number of characters in a text message is limited in some social nets and mobile text messages. The need for compacting the text becomes essential as it translates to saving online storage space, cost and many more factors. The paper proposes a text compaction system for Tamil, a first of its kind in Tamil. The system proposed in this paper handles common Tamil words, acronyms/abbreviations and numbers. Morphological analyzer [1] and Morphological generator are used to stem inflexion words and replace them to compact using a mapping repository. The proposed work is tested with over 10,000 words and it is found that the final result is reduced to 40% of the original text. The paper concludes by discussing possible extensions to this system.