A Hybrid VADER-SVM Framework for Enhanced Sentiment Classification

Abstract

Social media sites, especially Twitter, has become an essential avenue for passengers over the years. They present their thoughts regarding airline services. The analysis of unstructured textual data discusses the important insights about client happiness and sentiments. In this paper, an integrated supervised Support Vector Machine (SVM) classifier and a rule-based model called Valence Aware Dictionary and Sentiment Reasoner (VADER) are used as sentiment analysis technique to examine public opinion of major airline firms. The dataset comprising tweets of six American airlines is preprocessed using tokenization, case normalization, and stopword removal. The VADER model is used to classify tweets into positive, neutral, and negative attitudes for sentiment scores prediction. The effectiveness of VADER is proven by comparing the predicted values with the ground-truth labels. The word clouds are created to visualize sentiment-heavy phrases qualitatively. A SVM model is simultaneously trained using a feature extraction technique using TF-IDF, that yields dependable classification results. Various evaluation metrics such as categorization report and confusion matrix are used for quantitative assessment. Moreover, this paper presents airline recommendation using total sentiment score. It is observed that integrated rule-based and machine learning techniques compute reliable sentiment classification. The airline stakeholders have practical insights to improve customer service and engagement.