Real time multi feature performance analysis model for efficient prediction of student performance using neural networks

Abstract

The problem of student performance analysis and prediction has been well studied. There exist number of approaches to handle this issue which would use variety of features like the involvement in sports, education, academic and others. However, they suffer to achieve higher performance in predicting the performance of the student. To handle this issue, an efficient real time multi feature Performance Analysis Model (MFPAM) is presented in this article. The model consider number of features include the number of sessions the student attended, the number of seminars appeared, number of tests cleared, number of webinars accessed, number of assignments submitted, number of queries produced, number of subjects cleared in each time stamp, number of sports played, number of videos visited, number of extracurricular activities appeared and so on. By analyzing each features in different time stamp of the academic carrier, the method predict the performance of the student. The logs of student have been trained with neural network and at the test phase, the neurons perform analysis on each factor to produce support on different performance factors. At the output layer, the neurons generate set of weight measures towards various class of performance.