Exploring data mining possibilities on computer based problem solving data

Abstract

This article deals with the data mining process on a database containing the strategic steps made during complex problem solving in a computer-based assessment environment. The used database was constructed based on the log-files generated by the computer-based assessment software and published on-line by PISA. Preprocessing, attribute extraction and classification of the data by naïve Bayes classifier is presented. Evaluation parameter values such as the ROC, PR curves, the areas under these curves, as well as confusion matrices are given. Results indicated that the data contained in the database can be used for classification of the achievement in resolving the complex problems given in the PISA tasks.