Volume 18, No. 6, 2021
Sentiment Analysis Of Bengali Text Using Machine Learning: Novel Approach
Ms.Moumita Pal , Dr. Rajesh S. Prasad
Abstract
In the machine learning domain, sentiment analysis has emerged as a key framework for scientific and commercial market research. As there are few research works on sentiment analysis for this language, it is currently a more significant research field of Bangla language processing system. Sentiment analysis is essentially an automated text mining procedure that determines the emotion of a given text. A given text can be classified into many emotions using sentiment analysis. This paper focuses on sentiment analysis in the context of Bangla language. Supervised machine learning classifiers such as logical regressions, K-Nearest Neighbour (K-NN), linear supervised vector machine and random forest are applied to the feature matrices. Using the Linear SVM technique, the Unigram model had a precision of 83.26%at the dataset. While the Bigram model approaches accuracy at 72.04% and precision at 85.2%, the Trigram model has the highest precision score of all, at 92%.
Pages: 7117-7126
Keywords: K-NN, sentiment analysis, SVM, RF, LR, unigram, bigram, trigram.