Shooter video games for personality prediction using five factor model traits and machine learning

Video Games technology has rapidly increased in the last decade with increasing number of users. Players information becomes very important to the researchers in order to measure the impact of these Video Games on the players behaviors. First Person Shooter (FPS) game is considered one of the most attractive game, which make it a realistic environment to measure the players behaviors, emotions and interests. In this paper, we developed a custom First Person Shooter game called The Protector in order to study the impact of Video Games on the players' personalities. The game was published in the Google Play Store in order to collect custom data from users in un-controlled environments. The data was collected based on Five-Factor Model (FFM) for person-ality traits, which is considered one of the most well-known model in this field. Thousands of real user's data were collected using our published game. Then, different machine learning techniques have been used for the purpose of the personality classification and prediction. We used the classification task in order to automatically classify the possessing level of the players traits based on their data. We have used the prediction task in order to predict the player performance and behaviors in the future rounds. It is shown that the ANN algorithm provides the best accuracy of 98.6% over the other used algorithms in the classification task. While the SVR and KNNR provide the lowest error rate over the other used algorithms in the prediction task.

Automated summary

In this study, a custom First Person Shooter game called The Protector was developed to study the impact of Video Games on players' personalities. Using machine learning techniques, the researchers collected user data and classified their traits based on the Five-Factor Model. The study showed that the ANN algorithm had the highest accuracy in personality classification, while the SVR and KNNR algorithms performed best in predicting player performance and behaviors.