Data Pre-Processing: a Case Study in Predicting Student's Retention in Mooc

Abstract

Data pre-processing is a crucial phase prior to analytic task and yet rarely been discussed especially for e-learning data which has multilevel data. Providing a reliable data pre-processing is important to provide quality dataset. Therefore, this study investigates the problems arise in data pre-processing and in this case, for identifying the significant factors to implement prediction task. A MOOC dataset is selected for the data pre-processing task. The process in generating the summary of dataset is explained and the ultimate aim is to produce a dataset with features that are ready for data mining task. The study also proposed a process model and suggestions, which can be applied to support more comprehensible tools for educational domain who is the end user. Subsequently, the data pre-processing become more efficient for predicting student’s retention in MOOC.