This code applies Gradient Boosting on a continuous outcome variable (number of shares). Using a parameter search and two subsequent simulations, the best Parameter combination is approximated. The result outperforms Ordinary Least Square Regression in terms of Mean Absolut Error.