SnowLeopardR is an open-source machine learning-based anticheat project. It classfies player's "combat style" with the time series built from player's movements and interactions on entities.
The feature extration procedure is backed up with experiment results, and the subsequently derived dataset is testified to reveal user's movement characteristics. SnowLeopardR classifies the data by leveraging a learning vector quantization neural network[1].
The project is initially presented as SnowLeopard, which is more like a working draft. As a tutorial, it is also nice to have a bit of software engineering. Therefore, the original one has been refactored for a robust and scalable architecture and renamed to SnowLeopardR. For further information, please move to the link below.
[1] T. Kohonen (1995), "Learning vector quantization", in M.A. Arbib (ed.), The Handbook of Brain Theory and Neural Networks, Cambridge, MA: MIT Press, pp. 537–540
Java 8+
Maven 3
https://www.spigotmc.org/threads/machine-learning-killaura-detection-in-minecraft.301609/
GPL v3
Written with StackEdit.