Modeling and Predicting Scientific Thinking Skills of University Students Using a Data Mining Tool

Abstract

In order to model and predict scientific thinking skills of university students, scientific thinking test with psychological test and demographical information survey had been administered to students. Results obtained from the test and survey was analyzed to make pattern analysis. During preprocessing, outliers and invalid data were eliminated. Clustering methods, PCA analysis, regression and decision tree models were used during analyzing stage. As data analytic tools KNIME toolkit and Weka knowledge flow is used for further analysis and model generation. Socio-economic factors are found to have direct influences on the level of critical thinking