Prediction of the school administrators, who attended an action learning course, based on their conflict-handling styles: A data mining approach.
Author:
Abstract
Keywords
Abstract
In the recent years, school administrators often encounter various problems while teaching, counseling, promoting and providing other services which engender disagreements and interpersonal conflicts between students, the administrative staff and others. Action learning is an effective way to train school administrators in order to improve their conflict-handling styles. In this paper, the data mining techniques are used to determine school administrators who attended an action learning course based on their conflict-handling styles. To this end, ROCI-II instrument is used that consists of both the demographic information and the conflict handling styles of the school administrators. Various data mining techniques such as decision trees, discriminant analysis, support vector machines (SVM), k nearest neighbors (k-NN) and ensemble learner are used in prediction purposes. Various experimental works, on computer environment are carried out to validate the proposed idea. All data mining methodologies are simulated on MATLAB environment with 5-fold cross-validation technique. The classification performance is measured by accuracy criteria. The prediction experiments are conducted based on the two scenarios. In the first one, all statements (instrument items) are used to predict if a school administrator is educated or not with an action learning program. In the second scenario, the five independent dimensions are used individually to predict if a school administrator is educated or not with an action learning program. According to the obtained results, the first scenario achieves the best performance with k-NN method where the accuracy score is 75.0%. When the results from second scenario are considered, it is seen that integrating and compromising dimensions produce better accuracy scores than the other dimensions. Both integrating and compromising dimensions perform 73.7% accuracy scores with SVM and k-NN methods, respectively