From: Miguel R. C. <m....@la...> - 2008-05-20 14:21:50
|
Hi all, I finish branch /branches/miguelrojasch-IP which was supposed to close the bug predicting the necessary energy to ionize orbitals localized principally in heteroatoms and double bonds. The improvements has been: - Model to predict ionizatio potential based on weka software with the function LeastMedSq - Divided the classification in different models according with which group they belong - Removed bug 1860497 which Descriptors are not anymore dependents between them - Create new Interface for IChargeCalculator which makes clean the code grouping blablaCharges methods. IChargeCalculator is a extension of IElectronicPropertyCalculator - Added method in StructureResonanceGenerator which extracts container in resonance given an IAtom object - etc.. Here I also attached the junit-test-results comparing with trunk. It would be very nice if you could take at look before to submit to trunk and replay suggestions. Thanks in advance, Miquel |