Measurement of the appropriateness in career selection of the high school students by using data mining algorithms: A case study

Abstract

Less than optimal choice of the university department is one of the serious problems Turkish high school students have been suffering. There are a number of potential factors affecting the student's choice of her future profession. Some of these have received attention in the literature, but such studies do not always involve an investigation of the relationship between the factors analyzed and subsequent levels of academic achievement. The present study examines the relationship between the level of academic achievement and the students' abilities, interests and expectations, by using different data mining methods and classifiers, as a preliminary work to develop a system that will guide the student to selecting a career that will be a better match for her in the future. C4.5, SVM, Naive Bayes and MLP algorithms are used for the analysis; 10-fold cross validation and train-test validation are used as models to evaluate the classifiers results. The student feature set is obtained through questionnaires and psychometric tests. The questionnaire and the psychometric test were applied to 210 and 52 students respectively, from the Computer Engineering Department at Cumhuriyet University. The class was labeled either 'successful' or 'unsuccessful' with reference to the grades received by each student in computer engineering courses. The comparisons of various data mining algorithms, different data set results, and models used are presented and discussed.