A Magic the Gathering machine learning application. Currently in development with no intended release date. Application utilizes the Microsoft Kinect sdk version 7 and Microsoft Speech (Note: Not system.speech).
In order to have a more interactive experience for the user, the code needs to support "wildcards". Example, "I summon Chronomaton", should add a card to the field that will have the name "Chronomaton", as well as any other detail we can pick up from the audio stream. This is supported in System.Speech, but is not support in Microsoft.Speech for reasons unknown.
The only known variable about the Machine learning project, is that I would like to utilize the K-Means cluster algorithm. However, this will require an extensive amount of detail in order to determine who is winning a game, what is a possible "game changer" that can be made is, as well as any other problems. Look into why K-Means would be a viable algorithm, and the problems with it. As well, look into other possible algorithms that may solve these problems better than k-means.