refactor: save and select only max accuracy model

[CaptainJeff]

Updating the Xboost trainers to only save the best model instead of increasingly better outputs. Also adjusted the XGBoost_Runner.py to select the most accurate model from the available models

[bklynate]

@kyleskom - I am curious to know why the model is hardcoded instead of having it be dynamic and selecting the model with the highest accuracy.

[kyleskom]

When I first created this as a project to teach myself ml, that was never a concern to be and was just easy. As the project grew and became some what popular, I never went back to fixing that.

[XZoro-404]

From my understanding it is not always advantageous to use the highest percent model since that will lead to over fitting. So I really don't know on how to dynamically pick which model to use. That being said I am no pro at this so if someone has a good idea on how to implement this the correct way that would be great because I do agree that making it somewhat dynamic would be nice.

I am also not a huge fan of some of the formatting choices you made for XGBoost model files. granted I am just looking at them through github and not my IDE lol.