On-line drift reduction for portable electronic nose instrument in monitoring indoor formaldehyde

In this paper, a new on-line drift reduction method based on component correction method is proposed for using in a portable electronic nose (Enose) instrument which is used for real-time monitoring indoor formaldehyde. For a portable Enose instrument which is used indoors, the external conditions are changed from time to time, so the drift direction cannot be easily obtained by using component correction method. Support vector machines (SVM) is used for predicting baseline of each sensor in portable Enose instrument. According baseline prediction and correction, the drift direction is obtained, and reduced by using component correction. And then, another SVM is used for predicting the concentration of formaldehyde. The results show that the method works efficiently for a portable Enose instrument which is used for real-time monitoring indoor formaldehyde.