Investigation of ensemble methods in terms of statistics
TIMMS 2019 example
In this study, it is aimed to determine the factors affecting the mathematics achievements of eighth-grade students through trends in international mathematics and science study 2019 and compare the classification performances according to sample sizes and the number of independent variables of Bagging and Adaboost methods. Firstly, the most important factors affecting mathematics skills were obtained by using feature selection methods. Then, the performances of the methods were examined according to the sample size and the number of variables. As a result of the analysis carried out, no obvious difference was found between the performances of the methods according to the number of independent variables. On the other hand, the performances of methods of this study varied according to sample sizes, and it was seen that Bagging method showed better classification performance than Adaboost method in all sample sizes, and the performance of both methods approached each other when a sample of 3000 units was used.