Investigation of problematic technology use in university students with data mining methods

Although using digital technologies in all areas of life was thought to be a necessity in previous years, this is now seen as a requirement. For this reason, the necessity and desire to use digital technologies, even in simplest transactions, unfortunately may lead to problematic uses over time. Digital technologies, which make life much easier for users of all ages, from young people to old people, can have dangerous consequences if they are not used in a controlled, conscious, and regular manner. The aim of this thesis is to evaluate university students' problematic technology use with data mining methods. In the study, analyses were conducted by using data mining methods. In order to collect data, Demographic Information Form, İnternet Addiction Test Short Form, Social Media Disorder Scale, Phubbing Scale, Fear of Missing Out Scale and Smartphone Addiction Test Short Form were used. The participants consists of 1238 undergraduate students studying in different departments at a university located in the east of Turkey. Before the data analyses, the data was examined through the data preprocessing process. In the data analysis phase, while artificial neural networks, support vector machines, K-EYK nearest neighbor and random forest algorithms from data mining prediction models were used, the k-means algorithm from the clustering method was preferred among the descriptive models. Data mining analyzes were applied to each of the sub-dimensions of problematic technology use separately and then interpreted. According to the results of the study, the problematic technology use of university students were predicted with data mining algorithms. In addition, the best results were obtained with ANN and RO algorithms according to the correlation coefficient criterion in estimating problematic technology use. Internet addiction score was .811 with RO algorithm, social media disorder score was .805 with ANN and RO algorithms, digital game addiction score was .735 with RO algorithm, phubbing score was .759 with ANN algorithm, fear of missing out score was .556 with ANN algorithm and the smartphone addiction score was .793 with the ANN and RO algorithms. In the analysis made by the clustering method, the students were divided into three groups. Each group was determined as low, medium and high level according to problematic technology use. Although the levels of problematic technology use differed in the clustering method, it was concluded that there was no control over the use of digital technologies in all groups.