Comparison of Different Machine Learning Approaches to Text Classification

Naive Bayes, logistic regression, linear support vector machine, and deep neural networks are estimated and compared from the point of fulfilling text classification. The naive Bayes approach can be used in real-time with big datasets, but it assumes that each feature makes an independent and equal contribution to the outcome. Logistic regression is a linear model that is easy to implement in classification problems but has poor performance on nonlinear data. The support vector machine algorithm is a linear algorithm that performs well in higher dimensions, but it requires significant time to process. The main advantage of the deep neural network model over the other techniques is that it reduces the need for extensive feature extraction and selection and the distributed representation of words as features input into the network.