Found a bug in code used to train previous (gamma) NN. The results were biased towards higher energies. Previous statistics were low but now they have improved.
Network inputs (before and after):
![]() |
![]() |
Energy networks (before and after):
![]() |
![]() |
X correction networks (before and after):
![]() |
![]() |
Y correction networks (before and after):
![]() |
![]() |
![]() | Name | Last modified | Size | Description |
---|---|---|---|---|
![]() | Parent Directory | - | ||
![]() | NNParaGammaXYCorrect..> | 2015-01-09 17:56 | 113K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 130K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 50K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 49K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 46K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 49K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 51K | |
![]() | NNParaGammaEXYCorrec..> | 2015-01-09 17:56 | 50K | |
![]() | NNParaElectronXYCorr..> | 2015-01-09 17:56 | 130K | |