Model neuronskih mreža za predviđanje matematičke darovitosti u djece (CROSBI ID 174480)
Prilog u časopisu | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Pavleković, Margita ; Zekić-Sušac, Marijana ; Đurđević, Ivana
hrvatski
Model neuronskih mreža za predviđanje matematičke darovitosti u djece
The paper aims to model a neural network that will be able to detect mathematically gifted pupils in the fourth grade of elementary school. The input space consisted of variables describing five basic components of a child's mathematical gift identified in previous research, while the scientifically confirmed psychological evaluation of gift based on Raven's standard progressive matrices was used at the output. Three neural network models were tested on a Croatian dataset: multilayer perceptron, radial basis, and probabilistic network. The performance of models is measured by the average hit rate obtained on the test sample. The results show that the highest accuracy is produced by a radial basis neural network, which correctly recognizes all gifted children. Such high classification accuracy shows that neural networks have potential to serve as an effective intelligent decision support tool that will be able to assist teachers in detecting mathematically gifted children, especially in schools with a lack of psychologists.
matematička darovitost ; neuronske mreže ; višeslojni perceptron ; radijalno zasnovana funkcija ; probabilistička mreža ; Ravenove progresivne matrice ; učiteljska procjena darovitosti
nije evidentirano
engleski
A Neural Network Model for Predicting Children’s Mathematical Gift
The paper aims to model a neural network that will be able to detect mathematically gifted pupils in the fourth grade of elementary school. The input space consisted of variables describing five basic components of a child's mathematical gift identified in previous research, while the scientifically confirmed psychological evaluation of gift based on Raven's standard progressive matrices was used at the output. Three neural network models were tested on a Croatian dataset: multilayer perceptron, radial basis, and probabilistic network. The performance of models is measured by the average hit rate obtained on the test sample. The results show that the highest accuracy is produced by a radial basis neural network, which correctly recognizes all gifted children. Such high classification accuracy shows that neural networks have potential to serve as an effective intelligent decision support tool that will be able to assist teachers in detecting mathematically gifted children, especially in schools with a lack of psychologists.
mathematical gift ; neural networks ; multi-layer perceptron ; radial basis function ; probabilistic network ; Raven's standard progressive matrices ; teacher estimation of gift
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o izdanju
Povezanost rada
Informacijske i komunikacijske znanosti, Matematika, Pedagogija