The proposed work presents the detection of hate speech in social media. In this context, the text mining techniques are valuable. This, in turn, requires an accurate classification algorithm as well as text feature selection technique which works well with the sentiment classification. Therefore, a review of existing techniques is performed first. According to literature, amongst the techniques suggested by the researchers, the GI (Gini Index), DF (document frequency), POS (part of speech) tagging, and IG (information gain)are popular and frequently used techniques for sentiment classification. Thus, these techniques are chosen to implement the text feature selection. Further, the SVM (Support Vector Machine) and KNN (k-Nearest Neighbour) algorithms are applied to classify the selected features. The experimental outcomes show that the SVM is accurate and efficient algorithm for classifying measured features. Additionally, it is seen that the combination of SVM with the feature extraction techniques POS and IG is also time efficient. The proposed work may be extended by use of the prevalent algorithms.