Implementing Machine Learning on a big Data Engine for e-Learning

Due to high volume, velocity, and variety (referred to as "3V") of data production, the term "big data" has recently emerged. Because of such fast data growth, Data centers of educational institutions are seemingly exploding as well. Therefore, big data is making its way to the field of E-learning. Nowadays large educational institutions feel the urgent need to analyze this flood of data to provide students with higher quality of services. Despite the fact that there is huge volume and variety of data with high velocity of production, the existing traditional methods are unable to analyze such data. In other words, there is little done on implementing learning analytics on big data engines. The author created a scalable machine learning system on a cluster of three machines suitable for large dataset operations using big data engine and implemented an early warning process with machine learning techniques. The dataset was obtained from online and semionline courses from The University of Tehran. Many features like access time, history of access and history of actions of students were extracted and used for the classification task. A parallel distributed classification process was run to predict students' success and failures. It offers a fast, scalable and in-memory process with apache Spark. Eventually, the research compares the efficiency of proposed method with current and common methods'. The results reveal that our model had a high classification accuracy and the larger the dataset, the more efficient our process will be. It also discusses a new term "learning analytic" and "big data" that are completely new and worth reading. My paper contains worthy information about implementing ELearning data mining on big data engines.