Abstract:

Although Self Organizing Map is one of the most widely studied and applied neural networks it has not been adequately exploited to model and to analyze data of multisensor systems, in particolar in the field of chemical sensing. In this paper an example of how chemical sensor arrays can take advantage from the use of self organize map is illustrated and discussed. In particular that set of sensor arrays called electronic nose are considered. The paper presents a number of tools for the extraction of information from a trained SOM. It is important to point out that the methodology here outlined is valid for any kind of multisensor application also in fields very distant from the chemistry and the sensors.