Article Preview
Top1. Introduction
Researchers have indicated that most teachers are facing major challenges about effectively tracking the progress of their many students, and requiring more assistance to manage their classes, enhancing their teaching skills, and preparing suitable curriculum content (Coates, 2005; Hwang et al., 2008b, Macfadyen & Dawson, 2010). To provide well suggestions of instructional designs with appropriate learning content, it is important to analyse students' learning status as well as review the teachers' instructional methods. In the past decade, several methods or systems have been proposed to diagnose the learning problems of students for specified learning content, such as a small extent of concepts or subject units (Hwang, 2003a, Hwang et al., 2012; Panjaburees et al., 2013).
To maintain the teaching quality, uniform material or curriculum is generally applied to school education or professional skills education. The design of the commonly compulsory curriculum aims to have students with different background establish a common basic concept through the instruction with same material. It is expected that students achieve certain level on basic knowledge. However, the knowledge background of students is different, such that students will have different views, ideas and interpretations of the same material. Therefore, how to provide diagnostic mechanisms and effectively detect the learning weakness of students for assisting teachers in planning their teaching strategies has become an important issue.
Among the existing methods for analysing students' learning status, educational data mining has emerged as a focal research area in recent years (Merceron & Yacef, 2011; Baker & Yacef, 2010; Lin et al., 2013; Romero & Ventura, 2010). The main objective of educational data mining is to analyse the various types of educational data and to generate useful information to solve educational research issues. The successful applications of educational data mining include the adaptive learning system (Hsieh & Wang, 2010), diagnosing system (Chang et al., 2005; Hwang et al., 2008a; Merceron & Yacef, 2011; Chen & Chen, 2009; Tseng et al., 2010), recommendation and remedial learning system (Chu et al., 2010; Khribi et al., 2009; Romero et al., 2009; Hsu et al., 2010; Milicevic et al., 2011; Martinez-Maldonado et al., 2013), and learning management system (Graf, Kinshuk, & Liu, 2009; Liu et al., 2010; Macfadyen & Dawson, 2010). Martinez-Maldonado et al (2013) aimed to explore the potential of an enriched tabletop to automatically and unobtrusively capture data from collaborative interactions. They applied many data mining techniques to assist collaborative learning, e.g., clustering (Martinez-Maldonado et al., 2013) and sequential pattern mining (Perera et al., 2009). Moreover, Merceron and Yacef (2011) have successfully conducted association rule mining on Learning Management System (LMS). In our survey, it can be seen that there were less reference has applied change mining and association classification mining for teacher’s educational purpose.