Generation of rules for expert systems by statistical methods of fermentation data analysis
Reinhard Guthke, Burkhard Ludwig · Acta Biotechnologica · 1994
Abstract Five new methods for determining the relations between kinetic data of fermentations are described and applied to an industrial antibotic fermentation process. The input data for these method are the elements of the distance matrix dij, which quantify the sum of the deviation squares between the time dependent kinetics x (t) of the fermentation runs i and j. For each measurable or calculable kinetic state variable, one n x n distance matrix must be calculated where n is the number of fermentation runs. All methods compare these distance matrices by statistical or graph‐theoretical approaches. The algorithms obtained are universally applicable if enough kinetic data are available, especially from more than 10 comparable fermentation runs. The algorithms were developed for the use in knowledge acquisition modules of expert systems.