An Intelligent Prediction System of Students Academic Performance based on Deep Learning and FPSOPCNN

Abstract

This paper proposes a new method based on text extraction techniques for predicting student outcomes using cognitive computation. Predicting student academic achievement is most helpful in helping educators and learners improve their teaching and learning processes. This shows that these students have different experiences that influence their level of information capture in the classroom as they have the potential to use different lenses for training. This document provides a predictive examination of student academic performance in Tamil Nadu College in India during the academic year 2018 and 2019. First, this work applies statistical examination to gain insights from the data. Then, two datasets were obtained. The first dataset contains variables obtained before the beginning of the school year and the second includes study variables collected two months after the beginning of the semester. Convolution Neural Network and Fuzzy Particle Swarm Optimization Pulse Coupled Neural Network (FPSOPCNN) are designed to predict the end-of-year student performance for each dataset.