Application of Fuzzy c-Means Clustering for Polymer Data Mining for Making SAW Electronic Nose

Polymers provide chemical interfaces to electronic nose sensors for detection of volatile organic compounds. An electronic nose combines a properly selected set of sensors (array) with pattern recognition methods for generating information rich odor response patterns and extracting chemical identities. Each sensor in the array is functionalized by a different polymer for diversely selective sorption of target chemical analytes. Selection of an optimal set of polymers from a long list of potential polymers is crucial for cost-effective high performance development of electronic noses. In this paper we present an application of fuzzy c-means clustering on partition coefficient data available for target vapors and most prospective polymers for selection of minimum number of polymers that can create maximum discrimination between target chemical compounds. The selection method has been validated by simulating a polymer coated surface acoustic wave (SAW) sensor array for monitoring fish freshness.