Exploiting dependency relations for sentence level sentiment classification using SVM

In the sentiment analysis, finding the subjective clues itself is a challenging task. In this work, we propose a new approach, which employs Support Vector Machine (SVM) for classification, exploits the dependency relations in a dependency tree coupled with a large lexicon resource obtained from twitter to create a feature vector. The experiment shows a significant improvement over the baseline approaches and results are on par with existing methods in two-class classification.