SENTIMENT ANALYSIS OF PRE-SERVICE MATHEMATICS TEACHER THROUGH NAÏVE BAYES CLASSIFIER: THE CASE OF MATHEMATICAL ABSTRACTION PROBLEM

Mathematical abstraction as part of mathematical thinking process is an important and fundamental process in mathematics and its learning. Pre-service mathematics teachers' experiences and sentiments towards mathematical abstraction can contribute to the way they teach in the future. This study involved 67 Pre-service Mathematics Teachers at one of the Universities in Central Java Province who aimed to analyze their sentiments towards mathematical abstraction problems. The data collection technique used a questionnaire to reveal the Pre-service Mathematics Teacher's response to abstraction problems. Sentiment analysis is used to analyze the responses given which are categorized into positive, negative, or neutral. The technique used in the research is Naïve Bayes Classifier Multinomial. The classification results show 62.9% negative sentiment, 24.2% neutral sentiment, and 12.9% positive sentiment. In addition, the model evaluation results show an accuracy value of 66.7% which indicates the reliability of the model in classifying the sentiments expressed by Pre-service Mathematics Teachers towards mathematical abstraction problems. Pre-service Mathematics Teacher sentiment towards mathematical abstraction problems is dominated by negative sentiment. This shows that the process of mathematical abstraction is still considered a complicated and confusing process.