Developed Python script to extract comments data from Amazon and Official site. Performed NLP based Tokenization, Lemmatization, vectorization and processed data in Machine understandable language Have used VADERS, ROBERTA and BERT models to find the sentiment of the reviews and used the ratings on the source to check the accuracy. also used the textBlob library for processing textual data.
The proportion clearly shows that the Roberta Pretrained Model performs better than VADER.