CHUNKER BASED SENTIMENT ANALYSIS AND TENSE CLASSIFICATION FOR NEPALI TEXT

The article represents the Sentiment Analysis (SA) and Tense Classification using Skip gram model for theword to vector encoding on Nepali language. The experiment on SA for positive-negative classification iscarried out in two ways. In the first experiment the vector representation of each sentence is generated byusing Skip-gram model followed by the Multi-Layer Perceptron (MLP) classification and it is observed thatthe F1 score of 0.6486 is achieved for positive-negative classification with overall accuracy of 68%.Whereas in the second experiment the verb chunks are extracted using Nepali parser and carried out thesimilar experiment on the verb chunks. F1 scores of 0.6779 is observed for positive -negative classificationwith overall accuracy of 85%. Hence, Chunker based sentiment analysis is proven to be better thansentiment analysis using sentences.This paper also proposes using a skip-gram model to identify the tenses of Nepali sentences and verbs. Inthe third experiment, the vector representation of each sentence is generated by using Skip-gram modelfollowed by the Multi-Layer Perceptron (MLP)classification and it is observed that verb chunks had verylow overall accuracy of 53%. In the fourth experiment, conducted for Tense Classification using Sentencesresulted in improved efficiency with overall accuracy of 89%. Past tenses were identified and classifiedmore accurately than other tenses. Hence, sentence based tense classification is proven to be better thanverb Chunker based sentiment analysis.