Publication

Abstract : Opinions are statements that reflects people's perception or sentiment.In This paper we have done sentimental analysis using two major NLP models - VADER and RoBERTA using sentiment analysis.The study begins with an overview of RoBERTa regarding its pretraining and optimization enabling dynamic data collection. Similarly, the legal status of VADER is explained in terms of being based on dozens of opinions and predetermined rules. We are testing these models using two different dataset customers for evaluating the accuracy and precision of each model in predicting the sentiment of text. The results are analyzed and compared to know the problems and weakness of the models.These tests include training and refinement of RoBERTa's emotional data list model which uses transformer model to process input sequences, while VADER relies on pre-generated lexicon text.Different methods are used to compare the performance of two models and identify their strengths and weaknesses in different situations.