Feature correlation method for enhancing fermentation development: A comparison of quadratic regression with artificial neural networks

Abstract A systematic methodology for improving the efficiency of fermentation process development is presented. It consists of the “smallest composite” experimental design and subsequent feature correlation and analyses. The methodology is applied to a simulation of tylosin fermentation. By looking at the whole parameter space of the simulation, the interesting behavior of a neural net changing from lower order to higher order during training is examined. The results show that, different sizes of neural nets within a certain range give an equally good prediction by using the “stopping training” technique, while quadratic regressions are sensitive to the size of the data sets.

[1]  George E. P. Box.  Choice of Response Surface Design and Alphabetic Optimality. , 1982 .

[2]  Kurt Hornik,et al.  Multilayer feedforward networks are universal approximators , 1989, Neural Networks.

[3]  Stephen A. Billings,et al.  Non-linear system identification using neural networks , 1990 .

[4]  H. O. Hartley,et al.  Smallest Composite Designs for Quadratic Response Surfaces , 1959 .

[5]  P. Gray,et al.  Metabolic regulation in tylosin-producing Streptomyces fradiae: phosphate control of tylosin biosynthesis , 1981, Antimicrobial Agents and Chemotherapy.

[6]  Alan J. Miller,et al.  A review of some exchange algorithms for constructing discrete D-optimal designs , 1992 .

[7]  M. K. Fleming,et al.  Categorization of faces using unsupervised feature extraction , 1990, 1990 IJCNN International Joint Conference on Neural Networks.

[8]  J. A. Leonard,et al.  Radial basis function networks for classifying process faults , 1991, IEEE Control Systems.

[9]  N. V. Bhat,et al.  Use of neural nets for dynamic modeling and control of chemical process systems , 1990 .

[10]  W. Westlake,et al.  COMPOSITE DESIGNS BASED ON IRREGULAR FRACTIONS OF FACTORIALS. , 1965, Biometrics.

[11]  F. Girosi,et al.  Networks for approximation and learning , 1990, Proc. IEEE.

[12]  Nam-Ky Nguyen,et al.  Construction of Optimal Block Designs by Computer , 1994 .

[13]  J. S. Hunter,et al.  Multi-Factor Experimental Designs for Exploring Response Surfaces , 1957 .

[14]  George E. P. Box,et al.  Empirical Model‐Building and Response Surfaces , 1988 .

[15]  P. Gray,et al.  Metabolic regulation in tylosin-producing Streptomyces fradiae: regulatory role of adenylate nucleotide pool and enzymes involved in biosynthesis of tylonolide precursors , 1980, Antimicrobial Agents and Chemotherapy.

[16]  Murray Smith,et al.  Neural Networks for Statistical Modeling , 1993 .

[17]  George Cybenko,et al.  Approximation by superpositions of a sigmoidal function , 1992, Math. Control. Signals Syst..

[18]  W. J. Studden,et al.  Theory Of Optimal Experiments , 1972 .