|
|||||
|
|
Polynomial Neural Network for Linear and Non-linear Model Selection in Quantitative-Structure Activity Relationship Studies on the Internet |
|
| Authors: | I. V. Tetko T. I. Aksenova V. V. Volkovich T. N. Kasheva D. V. Filipov W. J. Welsh |
| Affiliation: | 1. Department of Biomedical Applications , Institute of Bioorganic and Petroleum Chemistry , Murmanskaya, 1, Kyiv-660, 253660, Ukraine;2. Laboratoire de Neuro-heuristique, Institut de Physiologie, Université de Lausanne , Rue du Bugnon 7, Lausanne, CH-1005, Switzerland;3. Institute of Applied System Analysis , prospekt Peremogy, 37, 252056, Kyiv, Ukraine;4. Department of Biomedical Applications , Institute of Bioorganic and Petroleum Chemistry , Murmanskaya, 1, Kyiv-660, 253660, Ukraine;5. Institute of Applied System Analysis , prospekt Peremogy, 37, 252056, Kyiv, Ukraine;6. Department of Chemistry and Center for Molecular Electronics , University of Missouri-St. Louis , St. Louis, MO, 63121 |
| Abstract: | Abstract This article presents a self-organising multilayered iterative algorithm that provides linear and non-linear polynomial regression models thus allowing the user to control the number and the power of the terms in the models. The accuracy of the algorithm is compared to the partial least squares (PLS) algorithm using fourteen data sets in quantitative-structure activity relationship studies. The calculated data show that the proposed method is able to select simple models characterized by a high prediction ability and thus provides a considerable interest in quantitative-structure activity relationship studies. The software is developed using client-server protocol (Java and C++ languages) and is available for world-wide users on the Web site of the authors. |
| Keywords: | Polynomial neural networks non-linear model selection Internet PLS cross-validation |