Evaluating the Effectiveness of English Speaking Conversation Practice for Communication Suitable for Students using the Random Forest Technique: A Case Study of Students Enrolled in English for Communication Course at Rajamangala University of Technology

Abstract

The objectives of this research are twofold: 1) to compare the effectiveness before and after using a dialogue-based English speaking skills training model for effective communication, and 2) to identify guidelines for adjusting the English speaking skills training model using the Random Forest technique suited to students enrolled in the English for communication course at Rajamangala University of Technology Tawan-ok. The methodology were 1) a dialogue-based English speaking skills assessment questionnaire, and 2) observation of English speaking behaviour for communication. The research findings indicated that the average scores before implementing the speaking skills training model were 3.11, which was considered good. After the training, the average scores increased to 4.30, indicating a significantly higher level of proficiency. The statistical test yielded a t-value of 15.871. Therefore, it could be concluded that the English speaking skills training model significantly enhanced students' English communication abilities at a significance level of .05. Regarding the guidelines for adjusting the English speaking skills training model using the Random Forest technique, the data classification results demonstrated an accuracy rate of 86.78%, a recall rate of 85.4%, and a precision rate of 83.13%. Additionally, 18 rules were identified as having significant correlations. The adjusted training model focused on various topics, including Greetings, Feelings and Emotions, Health, Asking for and Giving Directions, Weather, and University Life.