Advantages of Artificial Neural Networks (ANNs) as alternative modelling technique for data sets showing non-linear relationships using data from a galenical study on a solid dosage form

Artificial Neural Networks (ANN) methodology was used to assess experimenta l data from a tablet compression study showing highly non-linear relationsh ips (i.e. measurements of ejection forces) and compared to classical modell ing technique (i.e. Response Surface Methodology, RSM). These kinds of rela tionships are known to be difficult to model using classical methods. The a im of this investigation was to quantitatively describe the achieved degree of data fitting and predicting abilities of the developed models. The comp arison between the ANN and RSM was carried out both graphically and numeric ally. For comparing the goodness of fit, all data were used, whereas for th e goodness of prediction the data were split into a learning and a validati on data set. Better results were achieved for the model using ANN methodolo gy with regard to data fitting and predicting ability. All determined eject ion properties were mainly influenced by the concentration of magnesium ste arate and silica aerogel, whereas the other factors showed very much lower effects. Important relationships could be recognised from the ANN model onl y, whereas the RSM model ignored them. The ANN methodology represents a use ful alternative to classical modelling techniques when applied to variable data sets presenting non-linear relationships. (C) 1998 Elsevier Science B. V. All rights reserved.