Research Information System
Thesis Type: Postgraduate
Institution Of The Thesis: Gazi University, Eğitim Bilimleri Enstitüsü, BİLGİSAYAR VE ÖĞRETİM TEKNOLOJİLERİ EĞİTİMİ, Turkey
Approval Date: 2021
Thesis Language: Turkish
Student: EYÜP AKBULUT
Supervisor: Nursel Yalçın
Abstract:
Every digital tool used in the age we live in produces data. These data are collected and processed for various purposes. Discovering meaningful information and patterns in data stacks is called data mining, and the processing of data obtained from educational environments is called educational data mining. Considering the number of students, teachers and education stakeholders in our country, a huge amount of data emerges. In order to increase the quality of education, these data should be processed and data-based decision mechanisms should be used. This research was carried out to contribute to the field of educational data mining and to determine the factors affecting the educational processes in schools. Within the scope of the research, year-end monitoring and evaluation data of educational institutions that implement 714 special programs and projects under the General Directorate of Secondary Education were analyzed with education data mining techniques. These data were collected by the Ministry of National Education, General Directorate of Secondary Education.. Within the scope of the thesis, the data purified from the data within
ix
the scope of the school names, province names, district names and the Personal Data Protection Authority's data were shared with the researcher. In the study, decision tree, clustering, correlation, regression and text mining algorithms from educational data mining techniques were used. The cases where the results obtained from the algorithms are the same as the results of other algorithms were examined. This study were carried out with the free RapidMiner program. The general situation of the schools was evaluated by grouping according to geographical regions. The factors affecting academic achievement were tried to be determined with decision trees. The relationship between correlation and regression analysis and research topics, and the general profile of schools were determined by clustering. Written expressions in the data group were evaluated with text mining.According to the research, it has been observed that physical places and projects in schools support a low amount of academic achievement. It has been determined that teachers should not have projects and competition activities in their schools in order to be successful in competitions organized for teachers. Schools are divided into two groups and in the upper group; The rates of discipline are low, it has the documents such as clean school and successes in competitions, and the rate of student placement to undergraduate programs is high. In the subgroup The rates of discipline are high, it has not the documents such as clean school and successes in competitions, and the rate of student placement to undergraduate programs is low.
Keywords : Big data, disciplinary punishment, educational data mining, erasmus, e-twinning, decision trees, student success, rapidminer, data, data mining