In this Rpub we will discuss different approaches to the Santander's Kaggle competition.

A Kaggle competition where the objective was to classify the surface on which a robot was moving. The data is composed of measurements such as the orientation, the angular velocity or the linear velocity.

An optimization algorithm whose objective is the search of the best parameters given a cost function.

This function allows to compute the importance of each variable in a model. Its an alternative to Random Forest importance function. It is based on the paper that won the Heritage Provider Network Health Prize.