Prediction of Senior Secondary School Students’ Academic Performance using Hybrid Machine Learning Algorithm

Abstract

The immense growth of data in terms of volume, often presents an exciting challenge for the development of data analysis tools capable of detecting patterns in this data. Data mining has established itself as a discipline that contributes tools for data analysis, knowledge discovery, and autonomous decision-making across a wide variety of application domains. One of these application domains is the field of higher education. One of the key priorities of every higher education system is the review and improvement of educational systems in order to improve their services and satisfy the need of their students. Scholastic  performance in the academics of school students is a growing concern in institutes for higher education, more than ever as predictive models constructed using classic quantitative methods do not generate reliable findings due to enormous volumes of data, attribute correlation, missing values, and variable non-linearity. However, data mining approaches function admirably in these circumstances. This paper proposes a novel hybrid algorithm, CNN/LVQ3, for predicting student academic performance.