Comparison of Machine Learning Classification Method on Text-based Case in Twitter

Abstract

As Artificial Intelligence (AI) and Machine Learning (ML) gaining momentum on industry and academic field, a deeper understanding for AI and ML are highly required. One of the most popular sub-field in this field is text analysis. This paper will discuss the performance of classification methods for text-based data and give the best choices of classification method in term of accuracy and training time, so that will help ML enthusiast to build ML project that does not require high computational cost. This paper aimed to give recommendation to practitioner and academic about which classifier best for text classification. This paper will limit its study in supervised learning only. The tested algorithm will be Support Vector Machine, Logistic Regression, Naive Bayes, Random Forest, and K-Nearest Neighbor. To simplify the project, text will be labelled into single-label data, not multi-label. The test shows that SVM gives best result, in term of accuracy and training time among other methods.