electronic energy from AM1?

 I have a small question concerning the electronic energy parameter
 generated during a AM1 single point calculation.
 We have a dataset consisting of 30 compounds and we are trying to
 correlate physico-chemical parameters with the retention-time (k)
 on different HPLC-columns. We have generated around 30 parameters,
 AM1 outputs included. In order to see the influence on the retention
 time from different parameters we have performed a principal
 component analyses and a PLS regression model. The mode seems OK,
 with reasonable Qsq and good fitting. Our problem is that one of
 the more significant components is  explained mainly by the
 Electronic energy. My question therefore is:
 How should this parameter be interpreted? How is this parameter
 calculated and what does it describe?
 e-mail: j.nilsson $#at#$ farm.rug.nl