Teachers’ sentiments to Bikol MTB-MLE: Using sentiment analysis and text mining techniques

Abstract

Mother Tongue Based – Multilingual Education (MTB-MLE) is one of the quality programs of the K to 12 Law (RA 10533) that raises the country’s standard of education. The research study specifically looks into the sentiments of the teacher participants in the Bikol MTB-MLE from Grades 1, 2 and 3 public elementary schools by classifying their opinions using sentiment analysis and text mining techniques. Set of validated researchers made semi-structured survey questionnaire were gathered from the responses of 1,365 respondents from the province of Albay. Manual cleaning and R software package were employed in classifying both negative and positive responses and also facilitated the data manipulation and graphical representation. The information was tested using Naïve Bayes algorithm and obtained 88% accuracy. The results indicated success; however, it encourages similar researches possibly involving other mother tongue languages.