Data analysis of multi-sensor arrays

Array sensors present data that can be subjected to a variety of methods of multivariate analysis. The form of data analysis depends on the information required, whether identification of substances (classification) or the concentrations of given compounds. In theory there is much information in even a few sensors if they are properly matched to the matrix of analytes to be encountered. Pattern recognition methods based on principal components analysis, or artificial neural networks are the most popular methods presently being used. For sensor arrays to be accepted for a given analysis attention must be paid to proper validation.