ANALYZING MOVIE REVIEW SENTIMENTS ADVANCED MACHINE LEARNING AND NATURAL LANGUAGE PROCESSING METHODS
Unknown authors
Abstract
The field known as "sentiment analysis" examines the underlying attitudes and opinions expressed in texts.Data sentiment analysis is a powerful method for conveying the opinions of any individual, group, or community.Data submitted by customers used to be analysed using ML and NLP.The model's sentiment analysis ability is tested using the IMDB dataset of movie reviews using ML and DL approaches.This study uses IMDb-trained ML and DL models to identify reviews as good or negative.The methods explored are SVM, Logistic Regression, Bi-GRU, and LSTM.Interestingly, Bi-GRU placed first among the models with 98% accuracy, 98% precision, and 98% recall, demonstrating its effectiveness in sentiment classification.The paper also describes certain preprocessing activities such as tokenisation and TF-IDF, with an emphasis on their utilisation in improving model outcomes.Directions for future research will be to investigate more ensemble approaches and study the generalizability of the results to more different languages and cultures to expand the possibilities of sentiment analysis.