Гібридні методи машинного навчання для прогнозування успішності вступу до закладів вищої освіти

Abstract

In this paper, an experimental analysis of the effectiveness of machine learning methods for solving the binary classification task to predict the success of admission to a higher education institution was conducted. The results confirmed that the hybrid classifier based on the sequential application of a Probabilistic Neural Network and the stacking ensemble of four classifiers based on a Support Vector Machine with four different kernel functions and Logistic Regression as a meta-algorithm showed the highest accuracy and can be applied in practice.