Experimental results: Overall comparative accuracy across all considered approaches.

Our solution outperforms the five Machine Learning approaches we have considered within our experiments: around 4% more accuracy than Naive Bayes and around 6% more accuracy than SVM. Although Naive Bayes has performed very well, its results are boosted by overfitting the classes that had more data (e.g., ) at the expense of others, such as for which it achieved 0 precision and recall. Unlike Naive Bayes, our approach has performed fairly uniform and consistent across all classes.