Current Proceedings on Technology
Yazarlar: Nazir Hawi
Konular:-
Anahtar Kelimeler:Human computer interaction,Data mining,Classification,Bayesian belief networks,Internet addiction,Internet addiction test
Özet: For over a decade, Internet Addiction has been a very pressing problem, worldwide. The potential of data mining as an analytical tool for the accurate classification of Internet addiction level of 817 adolescents is explored. The investigation and full exploitation of the collected massive real-life data using five classifiers led to important findings. While the Bayesian belief networks achieved the highest classification accuracy rate (92.10%), Naïve Bayes ranked second (89.72%). Thorough exploration using Bayesian belief networks led to identifying item 6 of the Arabic version of the Internet Addiction Test as well as all other demographic factors as useless in determining adolescents’ addiction level. Furthermore, Bayesian belief networks classification of unknown but actual individual cases was excellent.
Dergi editörleri editör girişini kullanarak sisteme giriş yapabilirler. Editör girişi için tıklayınız.