本研究預測 Apex Legends 玩家的網路遊戲成癮、睡眠問題、廣泛性焦慮症發生的可能性。透過發放問卷的方式了解玩家們的身心疾病的狀況後，利用這些玩家填寫的量表分數，使用監督式機器學習的方式使用玩家們的遊戲數據做訓練，並預測發生上述身心症狀的可能性。本研究使用邏輯回歸、決策樹、隨機森林、XGBoost、Na?ve Bayes去進行預測，發現隨機森林在預測網路遊戲成癮、睡眠問題、廣泛性焦慮症時有良好的表現，且在評估分類器的效能上，AUC的分數落在0.7上下，表示這些分類器能夠用在預測這些疾病上。此外本研究也會使用皮爾森相關係數去比對遊戲數據內哪些會與遊戲成癮、睡眠問題、廣泛性焦慮症有顯著的相關性的特徵。最後分析玩家族群、遊玩平台、年紀等等與本研究所探討的疾病之間的關聯。;This research predicts the likelihood of Apex Legends players internet gaming disorder, sleep problems, and generalized anxiety disorder. By issuing questionnaires to understand the players’ physical and mental illnesses. Using their scale scores to train on their game statistic by supervised machine learning and predict the likelihood of having physical and mental disorder. It uses logistic regression, decision tree, random forest, XGBoost, and na?ve bayes to make predictions in this research. It found that random forest has good performance in predicting internet gaming disorder , sleep problem and generalized anxiety disorder, and the AUC scores are 0.7, this situation means that this classifier can predict these diseases well. In addition, this research will also use the Pearson correlation coefficient to find which features in the game statistic are significantly correlated with internet gaming disorder, sleep problem, and generalized anxiety disorder. Finally, it analyzes the relationship between players, game platform, players’ age, etc. and find the relationship about these diseases in this research.
