Sentiment Analysis Using Machine Learning and Recurrent Neural Networks
Chayanika Basak, Aaliyah Beg, Pooja Kumari, Ritu Rani, Arun Sharma, Amita Dev · 2022
Sentiment Analysis is an NLP problem dealing with the understanding of emotions and assigning the tag of negative or positive to the tweets. For this purpose we have used carefully picked classification models inclusive of Bernoulli NB which yields accuracy 78.25%, Multinomial NB which yields accuracy 78.34%, Support vector Machines which yields accuracy 78.25% and Logistic Regression with best accuracy 79.63% among the ML models and the DL model Recurrent Neural Networks which gave accuracy 81.45% and through in-depth analysis of our data set we have preprocessed it in such a way that the dataset is clean and set to be trained in the models without underfitting or overfitting.