Comparative Sentiment Analysis Using Difference Types of Machine Learning Algorithm

Abstract

The company is becoming digital-focused in today's world company. Companies are selling their goods and searching for input from customers. It becomes difficult to say that the product is good or not based on their feedback when all the user writes their review about that is an item. That's where it comes to deep learning. By using this, we can derive thoughts or emotions from the consumer's written text. This is a study of emotion. It may determines the review's emotional status. Our plan senses views from analysis by the user whether they are good or bad. We use algorithms such as SVM, Naive Bayes, and some approaches. We use the algorithm of Naive Bayes because we want to learn how often in the text words occur. And then we use SVM to define positive or negative terms. For our researching purpose, we use the Amazon consumer review data set, which was available online. Some methods that we are using for preprocessing and cleaned the document where just words are left. We trained our model so well with twenty-four thousand data. So, it will give us the best accuracy and we make this model with the best algorithm and after that, it gives the accuracy of 98.39%. This project will help us in real life when we are having trouble with product reviews. Our machine will help us to determine which review is good and which review is bad and make a category of a positive and negative review and saves our time.