stacking_regressor: Function to create the stacking model
In angy89/hyQSAR: Hybrid QSAR model

This function fits a stacking regression model from the prediction of the previous one

stacking_regressor(pred_tr = abs.alpha.power.res$pred_tr,
  pred_te = abs.alpha.power.res$pred_te,
  y_test = abs.alpha.power.res$y_test,
  y_train = abs.alpha.power.res$y_train)

`pred_tr`	is a list containing the vector of predictions for the training set
`pred_te`	is a list containing the vector of predictions for the test set
`y_test`	is the vector containing the response values for the test
`y_train`	is the vector containing the response values for the training

an list containing the forllowing objects,

`fit`	the fitted stacked model
`yhat_test`	vector of prediction on the test set
`yhat_train`	vector of prediction on the training set
`Metrics`	Named vector with internal and external validation metrics