scale_dat: Scale features
In ant-stephenson/lhc: LHC Higgs Boson Classification Challenge

This define a function to scale features of a matrix with reference to another matrix useful because you can normalise X_train, and apply the same transformation to X_test. Use with caution if either dataset contain extreme outliers.

1	scale_dat(X, ref, na.rm = FALSE, add.intercept = TRUE)

`X`	matrix of covariates.
`ref`	matrix of covariates from which to calculate mu and sd.
`na.rm`	a logical to indicate if NAs should be stripped in mean and standard deviation computations.
`add.intercept`	a logical to indicate if column of 1s should be added to the output (Intercept)