E-crypto learning and applying data mining in education sector to improve learner's performance

Education is being positively and innovatively enhanced through technology to create new strategies of delivering academic information both online and face to face. Online, Distance and Blended learning are the modern strategies that are and going to impact teaching approaches in academic institutions, relatively through introducing e-learning platforms which are designed as models used to collect and analyze meaningful information online for assessing, analyzing and presenting data patterns taking place among learners. This study proposes an e-learning requirement that focuses on mining student educational data with aim of exploring the behavior of learning, interaction with the platform from data generated by students in an education environment. The data is collected from the event logs of students for analysis to determine a model to extract their behavior and predictive patterns in an e learning environment. The techniques applied in this study are decision tree decision making process of filtering content of data, cluster analysis performing the organization of pattern grouping in collections, and random forest for creating prediction model that closely provides accuracy of student's success. Elliptic curve and SHA-256 cryptography method is applied to ensure security per the concerns of user confidentiality both for the instructors and student when performing authentication entries into the platform and during chat interactions. Mining of student's data would help in projecting student's performance and interaction activities, where the random forest technique surpasses logistic regression and support vector machine classification methods of machine learning predictions. With this approach learning institutions will be able to classify groups of learners with aim of analyzing and directing well informed decisions to help them improve in their learning sessions.