Predict students performance at any stage of degree program

Abstract

Data mining is the process of discovering new patterns from large data set using di erent data analysis techniques. During the last few years, the appli- cation of data mining techniques on educational data has gained importance. The education data mining helps in discovering hidden patterns related to various students academic activities and in predicting future performance based on existing data. With the growing number of software solutions for enhancing classroom environment, commonly called e-learning, collaborative learning or in general Learning Management System, the importance of edu- cation data mining is becoming more relevant for educational institutes. The objective of conducting this study is to analyze student s activity patterns and behaviors by applying data mining techniques and make predictions on their outcomes. Activity logs from Learning Management System (LMS) were collected for undergraduate students and were investigated through machine learning, data mining techniques and statistical models in an attempt to investigate how student activities, resource views, activities gap, previous semester grades, prerequisite course grades etc impact on the student per- formance. This research concludes that previous semester grades as well as rst term activities have highest impact on student grades. The outcome of this research will help in better understanding of how various parameters e ect students performance positively and negatively. Trends and patterns found using data mining techniques help the management in decision making process. The results might be employed to help students to get aware that in which course they need to focus to improve their performance, institutes design their courseware, assist the instructor to identify the students needing special attention and take desirable measures.