Still having troubles getting the positron ratio for runs with a raster. When I now include the raster I get
Before:
For a run with the raster on (73006) I get
Before:
I suspect that something funny is happening in the training. Need to re-check simulation/training scripts. The width at the target compared to the "no raster" reconstruction does not make sense. Improvement to the "energy correction" network could likely fix this.

New neural networks:

• Trained networks with larger raster and longer target in order to avoid biasing.
• Interestingly, there is a large impact on the cluster position and energy reconstuction on the cluster "Kurtosis". See training output. Not sure if this is a good or a bad thing. The Kurtosis is a measure of how peaked the distribution is. The Moliere radius could be too big in the simulation, ie, I need to raise the individual block threshold in the monte carlo.
• Raster "improved" neural networks still have problems.

To do:

• Compare 72994 (carbon without raster) to another carbon run with raster. Run 73006 is C+He with the raster on
• Train NN with enlarged raster/target Z size.
• Train new NN with reconstructed Z.

Other things:

• Asymmetries per run are looking good. For example
• Where can I find update values for these asymmetires per run? I am showing James's but I know that the packing fractions have been updated a few times since they were produced.
• New Polarized Radiative corrections code is close to complete. We nearly done debugging and performing various cross checks.
• Here is an example of the cross section pieces for 3He.

	Name	Last modified	Size	Description
	Parent Directory		-
	positron_id6_c2_7300..>	2015-01-09 17:56	1.0M
	positron_id2_73006_n..>	2015-01-09 17:56	32K
	positron_id2_73006.png	2015-01-09 17:56	33K
	positron_id2_72994_r..>	2015-01-09 17:56	31K
	positron_id2_72994_l..>	2015-01-09 17:56	30K
	positron_id2_72994_O..>	2015-01-09 17:56	30K
	positron_id2_72994.png	2015-01-09 17:56	31K
	dis_diff_xsec_nucleu..>	2015-01-09 17:56	15K
	asym_vs_run-30.png	2015-01-09 17:56	27K
	NNParaElectronXYCorr..>	2015-01-09 17:56	143K
	NNParaElectronXYCorr..>	2015-01-09 17:56	63K
	NNParaElectronXYCorr..>	2015-01-09 17:56	66K
	NNParaElectronAngleC..>	2015-01-09 17:56	41K