Comparing Supervised Machine Learning Strategies and Linguistic Features to Search for Very Negative Opinions

In this paper, we examine the performance of several classifiers in the process of searching for very negative opinions. More precisely, we do an empirical study that analyzes the influence of three types of linguistic features (n-grams, word embeddings, and polarity lexicons) and their combinations when they are used to feed different supervised machine learning classifiers: Naive Bayes (NB), Decision Tree (DT), and Support Vector Machine (SVM). The experiments we have carried out show that SVM clearly outperforms NB and DT in all datasets by taking into account all features individually as well as their combinations.

keywords: sentiment analysis, opinion mining, linguistic features, classification, very negative opinions