Tutorial Neural Networks

The usage of a neural network varies depending on the aim and especially on the network type. This tutorial covers two applications on the one hand the usage of a Kohonen network for classification, and on the other hand the prediction of object properties with a counter-propagation network, 

The following steps give a general outline for the usage of a neural network The nomenclature is that used in the web-based neural network included in the ELECTRAS system [9. For other neural networks the parameters may be slightly different. 

The structure coding depends strongly on the properties which arc to be modeled. For an explanation of the different structure codes and their applications. sec Chapter 8, 

Before the network training is launched various network parameters have to be chosen  

The different classes are identified by different coloi s and a map shows the classification. 

---

Neural networks were trained on the basis of these codes to predict chiralit> -dependent properties in enantioselective reactions [42] and in chiral chromatography [43]. A detailed description of the chirality codes is given in the Tutorial in Section 8,6,... 

This tutorial uses the MATLAB Control System Toolbox, the Fuzzy Logie Toolbox and the Neural Network Toolbox. Problems in Chapter 10 are used as design examples. 

B.J. Wythoff, Backpropagation neural networks — a tutorial. Chemom. Intell. Lab. Syst., 18 (1993) 115-155. 

Goodacre, R. Neal, M. I Kell, D. B. Quantitative analysis of multivariate data using artificial neural networks A tutorial review and applications to the deconvolution of pyrolysis mass spectra. Zbl. Bakt. 1996,284, 516-539. 

Wythoff BJ (1993) Backpropagation neural networks. A tutorial. Chemom Intell Lab Syst 18 115... 

There are many excellent introductory books and journal articles on the subject of neural networks. Just a few of them are listed below in the references. Additionally, there are tutorials online at various web sites. However, the applications of neural network techniques to problems in molecular biology and genome informatics are largely to be found in scientific journals and symposium proceedings. 

For a discussion of statistical methods and neural networks, see the book by Ripley (1996) and articles by Ripley (1993), Cheng (Cheng Titterington, 1994) and Sarle (1994). Warren Sarle maintains an excellent neural network FAQ (frequently asked questions) web page for the Usenet newsgroup comp.ai. neural-nets this web page contains many tutorial discussions and references to books, reviews, journal articles and other neural network resources. 

In this chapter we have provided a tutorial on neural networks, showing you how they operate, how and when they should be used, and the many chemical applications to which they have been applied. Here we wish to make several comments and observations, indicating some possible avenues for further work. 

Central Neural System BBS has an electronic bulletin board containing 26 Mbytes of files related to artifidal neural networks including simulation packages, d aos, source code, tutorials and other text. Most is suited to IBM PC compatible... 

Multivariate calibration has the largest number of applications of chemometric methods in routine analysis for instance it became a widely used technique in quantitative analy.sis of complex mixtures by IR or UV detectors, especially in food chemistry and environmental analytical chemistry. A number of textbooks, tutorials, and reviews " have been published in this field, and software is offered by the instrument manufacturers. Applications of multivariate calibration is very widespread, ranging for instance from the determination of moisture in mushrooms to research for non-invasive measurement techniques of glucose in human blood. The multivariate methods mostly applied are PLS, PCR. and recently also neural networks, Typical calibration models describe the relationship between a set of x-variables (UV or IR absorbances) and one y-variable (concentration of one substance) although it is possible to derive models that simultaneously predict a set of y-variables. 

---