Bilingual Lexicon Approach to English-Filipino Sentiment Analysis of Teaching Performance

Abstract

Abstract The aim of this study is to formulate a strategy that can possibly calculate teacher performance by analyzing textual feedback. Expressing textual responses in quantitative form like average sentiment rating can actually provide opportunities for administrators to see if the numerical ratings given complement that of the comments. Our approach was designed to enable processing bilingual textual data. Findings of this study shows that there is strong correlation between teaching performance actual mean rating and average sentiment rating. Furthermore, the approach employed obtained 86% accuracy indicating that it is an encouraging technique, capable of analyzing the students' textual responses. In future work, the use of POS tagging can be explored to improve sentiment analysis accuracy. Employing machine learning methods may also be considered to discover techniques and alternative approaches to sentiment classification.