การจำแนกระดับการติดเกมคอมพิวเตอร์ออนไลน์ของนักเรียนชั้นมัธยมศึกษาตอนต้นโดยใช้นิวรอลเน็ตเวิร์ก / สกุลทิพย์ ตุ่ยสิมา = Classification of online game addiction for students in secondary education (m.1-3) using neural networks
This research presents the classification of game addiction level in secondary school students (M.1-3) with a sample group of 33 students who play game in the residence daily. Data was collected during 18 May – 26 July 2011. The knowledge was synthesized using Multi-Layer Backpropagation Neural Networks and Decision Tree Algorithms, which are Supervised Learning Algorithms. The accuracy of the obtained model was tested by 10-fold Cross Validation approach. This research classifies computer games, based on their characteristic, into four categories: Long Term, Casual, Real Time and Turn Base. The experimental results revealed that classification of game addiction level using Multi-Layer Backpropagation Neural Networks Algorithm provided percentage of accuracy for Long Term Game, Turn Base Game, Casual Game, and Real Time Game as 95.50, 93.18, 89.42, and 87.91 ,respectively. Classification of game addiction level using Decision Tree Algorithm gave percentage of accuracy for Long Term Game,Turn Base Game, Casual Game and Real Time Game as 88.76, 90.91, 91.13 and 87.50 ,respectively. And the result of statistical analysis found that the addicted childrens play game at average 101.15 minutes per day, the average play time of the fanaticize group is 55.67 minutes per day, and that of the normal group is 52.46 minutes per day.