ABSTRACT
Abstract. An electronic nose (e-nose) data logger (Cyrano Sciences Inc, USA), comprising an array of thirty-two polymer carbon black composite sensors has been used to identify six species of bacteria, commonly associated with eye infections, over a range of concentrations in saline solutions. Readings were taken from the headspace of the samples by manually introducing the portable e-nose system into sterile glass vials containing a fixed volume of bacteria in suspension. After some data pre-processing, principal components analysis (PCA) and other exploratory techniques were used to investigate the clustering of the response vectors in multi-sensor space. Then, three supervised predictive classifiers, namely multi-layer perceptron (MLP), radial basis function (RBF), and Fuzzy ARTMAP, were used to identify the different bacteria. The optimal MLP network was found to classify correctly 97.3% of unknown bacteria types. The optimal RBF and Fuzzy ARTMAP algorithms were able to predict unknown bacteria with accuracies of 96.3% and 86.1%, respectively. A RBF network was able to discriminate between the six bacteria species even in the lowest state of concentration with 92.8% accuracy. These results show the potential application of neural networkbased e-noses for rapid screening and early detection of bacteria causing eye infections and the possible development of a Cyrano e-nose as a near-patient tool in primary medical care.
