Electronic Nose Coupled with Support Vector Machines for Rapid Discrimination of Black Tea According to the Quality Levels

Abstract

The dried black tea samples made from fermented leaves and different batches collected from tea plantation and factory were measured sequentially with electronic nose (E-nose) measurements and sensory analysis to discriminate the dry black tea samples according to the quality levels. The volatile patterns collected from the electronic nose were initially subjected to principal component analysis (PCA) and linear discriminant analysis (LDA) for clustering observation. Then, support vector machines (SVM) with different kernels (linear and radial basis function) was used in order to classify tea samples in three distinct quality levels namely, quality level one (Q1), quality level two (Q2), and quality level three (Q3). The results showed that SVM with radial basis function kernel provided good discrimination of tea samples regarding the quality levels. The overall correct classification of the three sensory quality levels was 98% showing the correct classification for Q1, Q2, and Q3 to be 96, 98, and 100%, respectively.