make relevance of datasets equivalent

[fjandrad]

Grid seems to contain twice as much all the other datasets together. We suspect this might be creating a bias towards fitting the grid. So relevance between datasets should be made equal in the begining with the possibility of arbitrarily select the relevance

[fjandrad]

To Define

Where should we include this behaviour. In the Section that stacks or in the part where we do least squares?