Title: Computational intelligence suport for quantum chemistry

  • Objective: create a program advising computational chemists which mehtods and basis sets to use
  • Participants: part of the OpenMol collaboration and COST project, on our side W. Duch, Karol Grudziński
  • Time:  since 1995
  • Results: small prototype made by Karol, graphical interface made in Munich; several talks given.
  • Open problems: collect the data! We are ready to extract rules + analyse the data to find the best reccomendation method; best data should be based on real accuracies of calculations of molecular properties, but at the beginning recommnedations of a good expert on a large number of molecular calculations should be sufficient. What is the best selection of features?

    •  
  • Working Log (local access only)

    •  
     Waiting for the data Rafał Adamczak analyzed some QSAR data.