Counteraction of environmental disturbances of electronic nose data by independent component analysis

Electronic nose data are usually affected by several disturbances in part due to the non-idealities of sensors and in part due to the effects of the environment as far as the composition of the gaseous sample is concerned. In some practical applications, the sample cannot be completely insulated from the surrounding environment and its drift (due to temperature and relative humidity changes) may be so large to completely dim the intrinsic resolution of the sensor array. The problem could be solved at data analysis level making some assumptions on the nature of the disturbances. s paper, the independent component analysis (ICA) method will be illustrated and used to segregate the environmental disturbances from the meaningful part of the data. The method is tested on a practical and simple experiment aimed at distinguishing two classes of fruits measured for 1 week at different room temperatures and relative humidity. The application of ICA results in a meaningful improvement of the classification. The application of ICA gives new insights in the comprehension of the working mechanisms of electronic nose systems. Connections between ICA and electronic nose will also be discussed in the paper.