Comparison of Parameters of Sentimental Analysis Using Different Classifiers

Abstract

Nowadays, we are seeing the surge of using social media as a platform for promoting and for any target. So, understanding the precise behavior of groups or individuals using his/her tweets or comments is the next step of sentiment analysis. We see innumerable knowledge shared on social media daily. We work on both sides on one side. We see various availability of data or opinions, and on the other hand, we challenge to group them in one centroid or domain in this work. We used the data of sentiment140 from Stanford University to perform various machine learning classifiers on feature extractors. Classifiers used in works are ridge classifiers, logistic regression, SVM, perceptron, passive aggressive classifiers, stochastic gradient descent, Naive Bayes, KNN, nearest centroid, and adaptive boost classifier. Our goal is a performance comparison of these classifiers on four performance factors namely recall, precision, accuracy, and F1 score. This work will help in consideration of selecting classifiers during sentimental analysis.