RMSE seems to be a popular metric but now one has to calculate it through np.sqrt(mean_squared_error(XXX, XXX)). Maybe we can add squared option to mean_squared_error and add a scorer neg_root_mean_squared_error.
Wiki page: https://en.wikipedia.org/wiki/Root-mean-square_deviation
RMSE seems to be a popular metric but now one has to calculate it through
np.sqrt(mean_squared_error(XXX, XXX)). Maybe we can addsquaredoption tomean_squared_errorand add a scorerneg_root_mean_squared_error.Wiki page: https://en.wikipedia.org/wiki/Root-mean-square_deviation