Developing Automatic POS Tagging and Summarizing Techniques for a Low Resourced Language

Abstract

Language has always been a powerful source of communication for human. It has newlinebeen a medium of expressing thoughts, emotions and feelings. Languages have gone newlinethrough various phases, from its primitive form to modern languages. The present human newlinehas brought it to a stage that can be processed by a machine and the entire activity is newlinebaptized as natural language processing (NLP). It is a field that is an integration of newlinelinguistic knowledge and artificial intelligence. It comprises various subfields like speech newlineprocessing, text to speech conversion, sentiment analysis, machine translation, Parts of newlinespeech (POS) tagging and automatic text summarization (ATS). As languages are varied newlinetheir breakthrough in the digital world has attained different heights. The variations in the newlinescript, the ambiguous nature of the words, morphological complexity, learning the context newlineand accordingly POS tagging the words and extracting the semantic are some of the newlinechallenges faced by NLP. European languages like English have made great strides in the newlinecomputational platform but there are some languages that are not backed by resources that newlinecan enable them to be processed by a machine. For example Indian languages particularly newlineOdia is far behind in the race. Its morphological complexity and lack of machine readable newlinecorpus has stood as a barrier in the progress. The existing NLP tools cannot be newlineimplemented directly on Odia language. So this research work aims to design efficient newlinesystems that solves two use cases POS tagging and automatic text summarization and give newlinean optimized result. It conducts the experiment considering Odia language to bring it into a newlinecomputational platform. newlineAutomatic text summarization has marked its importance with the rapid growth of newlineinformation in this digital world. The urge to access the knowledge base in limited time newlinehas prompted the researchers towards developing automatic text summarizers that can newlineextract the synopsis from the crowded data. The present research work makes a positive newlineattempt to solve the issue