This is an ongoing analysis on choosing the best model through for the African Soil Prediction Kaggle Competition.
Update 1 (08/20/2019): The support vector machine model definitely shows better performance over using a gradient boost regressor. I will continue to go down this route, increasing the number of iterations over time as well as continuing to use the cross validation scoring function to tally scores.