In gas sensing applications we often use sensors which have poor selectivity. Such sensors are mature, cheap devices with simple application circuits. Unfortunately, their cross sensitivity greatly restricts their usefulness when used in systems which employs univariate analysis. One method which allows to overcome this difficulty is utilization of multivariate analysis methods applied to the response from the group of such non-selective sensors. This work presents the multivariate, pattern recognition system utilized to process the response of an array of non-selective semiconductor sensors to obtain the qualitative information.
The system composed of the array of eight semiconductor, chemoresistive gas sensors was used for the classification of hydrogen, methane and carbon oxide gaseous samples. The classification task was performed by pattern recognition methods applied to the multivariate response of the array. The pattern recognition scheme used for classification uses a feature subset selection algorithm coupled with an objective function designed for clustering and a multilayer perceptron classifier.
For an array of eight chemoresistive gas sensors a computer pattern recognition system was built. Multivariate data analysis was performed for measurements of three gases' dilutions: hydrogen (H2), methane (CH4), and carbon monoxide (CO). The pattern recognition system included a feature subset selection algorithm involving PCA and objective function. Dimensionality reduction was applied to two kinds of patterns: three aforementioned gases and six dierent concentrations of hydrogen. For patterns of the three gases, classication tests were performed using k-NN algorithm and N -fold based validation method.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.