Izvorni znanstveni članak

DETERMINING DOMINANT FACTOR FOR STUDENTS PERFORMANCE PREDICTION BY USING DATA MINING CLASSIFICATION ALGORITHMS

Edin Osmanbegović ; Faculty of Economics in Tuzla
Mirza Suljić ; Faculty of Economics in Tuzla
Hariz Agić ; Prosvjetni zavod, Tuzla

Sažetak

The central problem in the process of a discovering knowledge from data, in the field of educational data mining, is to identify a representative set of data, on whose basis a classification model will be constructed. This paper presents the research results in reduction of data dimensionality, in the classification problems of prediction of student’s performances on the example from high schools, in Canton Tuzla. In this paper are shown different algorithms that are
used to reduce the dimensionality of data and to development of a data mining model for predictions of performances of students, on the basis of their personal demographic and societal features. It was found that algorithms Random Forest and J48 generate classification model with and accuracy higher than 71%.

Ključne riječi

data mining; educational data mining; predicting performance; student success; secondary schools

Hrčak ID:

133734

URI

https://hrcak.srce.hr/133734

Datum izdavanja:

2.2.2015.

Posjeta: 3.422 *