Machine Learning Based Strategy for Students Performance Improvement

Performance prediction is a tool that may help to understand the students' behaviors and challenges for their overall academic growth. Though the whole teaching pedagogy has enormous hills and valleys and cannot be clubbed together, some data mining and machine learning studies can be done to recapitulate the students' clusters according to their mindsets and capabilities. Student result prediction is tremendously important because student performance is the first step that helps in selecting electives for their advanced learning and placements too. The internal exams are the midterm evaluations of the student that help a teacher in understanding the level of individual students in the class. These marks obtained by the students as internal marks can forecast their result in the final exam having the highest weightage of the subject. In this proposed work, authors are figuring out the student's probability of passing the exam using a logistic regression-based machine learning algorithm. This work may help recommend the formation of different clusters for slow, average, and high-level learners, so that teachers may plan their attention according to the exact caliber of their audience. The study can be implemented in any centralized examination system to predict the final results based on the internal classroom performance exams. This study is carried out by proposing a sigmoid function-based hypothesis solved with logistic regression with three performance exams. These performance exams imply the development of three input variables-based algorithms for the proposed study.