DEVELOPMENT OF THE LINE CHATBOT “SPARKISTICS” TO ASSIST TEACHERS IN ANALYZING RESPONSES AND RECOMMENDING METHODS TO ENHANCE MATHEMATICAL REPRESENTATION COMPETENCY IN QUALITATIVE DATA ANALYSIS AND PRESENTATION

Abstract

This research resulted in the development of the LINE chatbot to assist teachers in analyzing students’ responses and recommending effective ways to enhance mathematical representation competency. The LINE chatbot provides individualized feedback and tailored instructional guidance aligned precisely with each student’s answer, thereby enabling teachers to support student learning more precisely and in greater depth. The tools employed in this research included the LINE Official Account, LINE Developer, and Dialogflow, all of which are capable of advanced natural language processing. The development data comprised 148 sets of student responses from assessments on the topic Qualitative Data Analysis and Presentation, collected from Matthayomsuksa 6 students. Of these, 109 sets were used to train the chatbot and 39 sets were used to test its overall performance, along with an associated scoring rubric. Data analysis was conducted using a confusion matrix, average values, and percentages. The research findings revealed that: 1) the LINE chatbot developed in this research, named “Sparkistics”, has the following capabilities: 1.1) facilitating the seamless delivery of important documents related to the assessment, 1.2) analyzing students’ responses through interactive, word-by-word chat interactions, and 1.3) providing concrete and actionable recommendations for promoting students’ mathematical representation competency; and 2) the chatbot’s performance in analyzing responses and assigning proficiency levels showed an average accuracy of 92.36%, precision of 92.36%, recall of 92.71%, and F1-score of 92.36%. The findings indicate that the “Sparkistics” LINE chatbot has sufficient potential for practical application to enhance mathematics teaching and learning in the classroom.