Analysis of students' incidents in higher education using data mining techniques.

摘要

Institutions of higher educational are the most important environments in which students, families, educators and community members have opportunities to learn, teach, and grow. However, one of the most problems that face the IHE's is the incidents of students' behavior. The objective of this study is to decrease the incidents of students' behavior by identifying the factors which cause the incidents in college campuses.;CRISP-DM Methodology has been applied to manage the process of data mining, four data mining techniques: J48 Decision Tree (DT), Naive Bayesian (NB), Artificial Neural Network (ANN), and Multinomial Logistic Regression (MLR) have been used to build the classification models and to generate rules to classify and predict the student's behavior and the location of incident in college campuses which will take into consideration seven factors: Student Academic Major, Student Level, Gender, GPA Cumulative, Local Address, Student Ethnicity, and time of incident by month.;Finally, all techniques were evaluated and compared. However, based on the evaluation and comparison it was found that the results of the accuracy were high for all the classification models; Multinomial Logistic Regression gave the highest accuracy, second was J48 Decision Tree algorithm, third was Artificial Neural Network, and lastly was Naive Bayesian Classifier.