Application of Cascade Correlation Networks for Structures toChemistry

  • Authors:
  • Anna Maria Bianucci;Alessio Micheli;Alessandro Sperduti;Antonina Starita

  • Affiliations:
  • Dipartimento di Scienze Farmaceutiche, Via Bonanno 6, 56126, Pisa, Italy;Dipartimento di Informatica, Università di Pisa, Corso Italia, 40, 56125, Pisa, Italy;Dipartimento di Informatica, Università di Pisa, Corso Italia, 40, 56125, Pisa, Italy;Dipartimento di Informatica, Università di Pisa, Corso Italia, 40, 56125, Pisa, Italy

  • Venue:
  • Applied Intelligence
  • Year:
  • 2000

Quantified Score

Hi-index 0.00

Visualization

Abstract

We present the application of Cascade Correlation forstructures to QSPR (quantitative structure-property relationships)and QSAR (quantitative structure-activity relationships) analysis.Cascade Correlation for structures is a neural network model recentlyproposed for the processing of structured data. This allows thedirect treatment of chemical compounds as labeled trees, whichconstitutes a novel approach to QSPR/QSAR. We report theresults obtained for QSPR on Alkanes (predicting the boiling point)and QSAR of a class of Benzodiazepines. Our approach comparesfavorably versus the traditional QSAR treatment based on equationsand it is competitive with ‘ad hoc’ MLPs for the QSPRproblem.