Akademska digitalna zbirka SLovenije - logo
ALL libraries (COBIB.SI union bibliographic/catalogue database)
  • A comparative study of polynomial and neural network modelling for the optimisation of clear coat formulations
    Tušar, Livija ; Tušar, Marjan ; Leskovšek, Nevenka
    The major purposes of this work were to acquire optimal recipes of polyestermelamine clear coats and to compare the results of two different tipe of model: neural network models and polynomial ... models. The process of optimisation was as follows: selection of the dependent and the independent variables, selection of the experimental design, measurements, optimisation of the architecture of the neural network, the neural network and polynomial modelling, evaluation of results, graphical representation of results and determination of the optimal recipe. The architecture of the neural network model (the number of neurones in the hidden layer and the number of iterations) was also optimised. The degrees of freedom for the statistical evaluation of neural network model were calculated on the basis of number of input, output and hidden neurones. Results of the neural network and of mathematical models were compared. It has been established that neural networks provide useful tools for modelling and for the prediction of the results together with the polynomial models.
    Source: Surface coatings international. - ISSN 1356-0751 (Vol. 78, no. 10, 1995, str. 427-434)
    Type of material - article, component part
    Publish date - 1995
    Language - english
    COBISS.SI-ID - 42070

source: Surface coatings international. - ISSN 1356-0751 (Vol. 78, no. 10, 1995, str. 427-434)
loading ...
loading ...
loading ...