Optimization of Electronic Nose Sensor Array for Tea Aroma Detecting based on Correlation Coefficient and Cluster Analysis

Abstract The electronic nose system is widely used to detect tea aroma, and the sensor array plays a fundamental role in detecting results. In this paper, we propose a sensor array optimization method based on correlation coefficient and cluster analysis. First, redundant sensors are assessed according to the correlation coefficient calculated between two sensors. The discriminating performance value (DPV) of the sensor is calculated, which reflects both the inter- and intra-class dispersion of tea varieties discriminated by the sensors. Thus, the redundant sensor with a small DPV is removed. Second, based on the cluster analysis (CA), the independence between sensors is analyzed. Beginning with the sensor with the highest DPV, the optimized sensor array can be acquired by iteratively selecting a candidate sensor that has the largest clustering coefficient with previous selected sensors. According to the above methods, sensor arrays for green tea (LG), fried green tea (LF) and baked green tea (LB) are constructed. Validated experiments are carried out by detecting 12 kinds of green tea by LG, 6 kinds of fried green tea by LF and 6 kinds of roasted green tea by LB. The classification accuracy using methods of Linear Discriminant Analysis (LDA) based on the average value (LDA-ave) combined with nearest-neighbor classifier (NNC) can almost reach to 100%. When used to discriminate between various grades of West Lake Longjingtea, LF shows better performance than that of the German PEN2 electronic nose.