Fuzzy-Assisted PLS Regression for Enhancing Quantification Efficiency of Electronic Nose

Electronic nose is a chemical sensor array-based chemical identification system for chemical analyte recognition and quantification. Regarding chemical analyte recognition methods, plenty of work is reported in the literature but relatively less attention is paid towards the quantification techniques. PLS regression is a well-known quantification technique. This work presents a fuzzy c-means clustering-based uncertainty measure computed by using the concept of Shannon entropy, for the data sample weighting, which enhances the quantification efficiency of PLS regression technique. The method is validated on the three electronic nose data sets available from the published literature.