Evaluation of Human Sentiment with Supervised Machine Learning Techniques

The estimation of human emotions in text is a vital task in many applications today starting from stock market prediction to social media surveillance. The following paper discusses some of the supervised machine learning approaches to sentiment analysis and focuses on their approaches, performance, and usage. This paper gives a gist of Bayes, SVM, Decision trees, Random forests, and ANNs as techniques. Moreover, we explain the problems associated with SIA, such as an uneven number of samples, selected attributes, and language differences’ effect on model accuracy.