An Affective Instruction for Mathematics Teaching based on Hybrid Kansei Engineering Approach

Abstract

One of the factors that influence the effectiveness of mathematics teaching is the design of mathematics course based on student's learning style. Based on their sensory preferences, students tend to have one or more of the three learning styles: visual, auditory, and kinesthetic. A traditional way to study and group students’ learning style is by asking students to fill in a questionnaire which may contain more than 100 questions. However, such an approach often leads to students’ boredom. To address this problem, we offer a new approach to group learning style by using opinion mining. Based on its motivation, the main objective of this research is to construct a mathematics teaching using Hybrid Kansei Engineering Approach. In particular, this research is focused on achieving the following objectives (1) To identify and group student’s learning style, (2) To verify and validate a mathematics teaching design. In this study, we consider 162 student opinions on mathematics teaching from three inclusive junior high schools in Bogor by spreading a one-page questionnaire. To study the opinions that can later be used to group the learning style and construct a mathematics teaching design, we have integrated three methods, namely Term Frequency-Inverse Document Frequency (TF-IDF), Semantic Differential (SD), and Principal Component Analysis (PCA). First, TFIDF is used to determine how important the words or opinions is to the collection. Then, SD Scales are used to evaluate student preferences. To extract the opinions as well as to cluster the learning styles, we use PCA approach, and Quantification Theory Type 1 (QT1) used to synthesize the relationship between teaching design and student preferences. The result of this study shows that 34% or 55 students had a visual learning style. Meanwhile, 12% or 20 students had an auditory style, and 54% or 87 students had a kinesthetic style. The results of this research indicate that the proposed approach can effectively group the student's learning style in mathematics teaching. Furthermore, it can potentially be recommended for teachers to design an effective mathematics teaching course. From these models, design support information was generated to assist teachers in developing the effective mathematics teaching. The result presented that all design specification of mathematics teaching was specific for each learning style and most of the rules were productive as the new design rules.