Towards the Development of Intelligent Agent for Novice C/C++ Programmers through Affective Analysis of Event Logs

Abstract

Novice programmers are having a difficult time writing their first programs and instructors handling a programming class with a large population are also having difficulties in monitoring every student's activity. One way to address this is to develop a tool that can help instructors monitor students' negative affects such that these instructors can easily determine students who need attention and guidance. This study established a model that is used in the development of an affective intelligent agent that can determine two negative affect states of students while making a program in C++, namely, confused and boredom. The model was established through the gathered key log data from novice C++ programmers. Processing these data through a data mining tool called WEKA, some patterns in detecting the affect states of these programmers were discovered. On the other hand, the affective intelligent agent can be used to improve monitoring student's attitude towards programming by alerting the teacher whenever the agent detects sign of boredom or confusion from the students.