Start optimisation / solve least squares problem direct. Are we stuck in a local minima?
Don't necessarily need to benchmark produced charges, but would be good to see how different they are.
Ethers seem to be the most problematic. Need to analyse this further.
Perform analysis of the gradients during fitting - are ethers trying to constrain the parameters differently than the rest of the data?
Split the analysis into polar and non-polar. Electric field expected to yield better results for polar molecules over non-polar. Need to determine if the test set is balanced too much towards non-polar.
Potentially look more into fitting to ESP rather than electric field?
Maybe we need to expand the sets to include some nitrogen data to see the real benefits of the refits?
OM - Adding N data would be large expansion of the scope. Need to rethink data set, number of BCCs involving N very large. Benefit of looking at smaller set is allows to tease out more easily the subtle effects going on.
Action items
Add action items to close the loop on open questions or discussion topics:
Owen Madin: Restart the BCC optimisations from randomly perturbed AM1BCC parameters.
Analyse whether the alcohol issue was resolved by the BCC optimisation.
Optimising the BCCs against phys-prop data (both with and without QM regularisation)
Analysis of ethers performance (especially looking at gradients).
Comparing the charges produced by the BCC study with the GAFF study.
Decisions
Type /decision to record the decisions you make in this meeting: