N-Gram and TF-IDF for Feature Extraction on Opinion Mining of Tweets with SVM Classifier

This work evaluates the performance of the Support Vector Machine (SVM) classifier on tweets opinion mining in five datasets available on the literature. For feature extraction, the N-Gram and TF-IDF, k-folds cross-validation techniques were used in the classifier modeling step. Variations of N-Gram with L-gram, 2-gram, and 3-gram combined with k-folds cross-validation in 10-folds, 15-folds and 20-folds yielded 63.93% to 81.06% accuracy. Satisfactory results were obtained, which can be improved with the application of an optimization technique to adjust the classifier parameters.