Eureka! Institutional Repository

Machine Learning (ML) algorithms and Virtual Reality (VR) technologies have received increasing attention in recent years in the field of IT. Machine Learning offers solutions to complex problems that require large computational power. It is widely used in predicting outcomes given some large data sets. Virtual Reality contributes to both entertainment and other fields such as industry and education. In this paper, we analyze techniques related to these technologies, and develop MM models for predicting movements with VR device controllers in a game. Both MM and Neural Network algorithms are compared, models are created for each of them, and their performance with respect to the binary classification problem is compared. For the purposes of the paper, a game was developed in the Unity engine using C#, which was used both to collect the model training data and to demonstrate its capabilities after training. The training was performed using the Python language and several of its libraries, including Keras, for Neural Networks. The result of the project was several trained models with high performance accuracy (up to 92.5% on new data), and a game in Unity for VR consoles.