Establish a Baseline
Elect a machine
Elect an evaluation criteria
Determine appropriate normalization
Train machine
audit for bias and batch effect
Return to normalization step