Linear Regression and Support Vector Machine for Classification of E-Learning Students’ Engagement and Performance

Dealing with continuous (frequently occurring) and huge volume data with conventional computing environment becomes challenging now days due to the memory and processing limitations present in the allocated resources though the network connectivity and site processing power are relatively high. However distributed processing approaches support this issue as a main theme through filtering data to our needs and analyzing the data to check the presence of required characteristics becomes solvable in specific contexts and IT industries thrives on this capability. In this paper the environment for e-Learning is selected as there are many inherent problems to be solved and researchers progress mainly due to technology advancement by using relevant tools. Here we address the problem of filtering or extracting the data not from any data warehouse but continuous data collection from connected nodes and isolated tables and generate the data for checking the relationships from inputs of student engagement activities to their performance. Since the inputs are not independent, it is applied with various filters through the queries and the patterns are detected by machine learning techniques like linear regression and support vector machine.