Classification of Green Tea Samples Based on Epigallocatechin-3-Gallate Content Using Molecular Imprinting Polymerization Technique

The proposed system implements a machine learning based approach to classify samples of green tea based on the EGCG (Epigallocatechin-3-Gallate) content. The EGCG content in green tea samples are estimated by using the synthesized molecularly imprinted polymer (MIP) electrode. The voltametric responses were obtained using a three-electrode set-up. The data was recorded and pre-processed. The responses obtained after pre-processing are fed for feature selection using UMAP (Uniform Manifold Approximation and Projection) algorithm, in which the number of features is reduced to generate optimized output. Further, the selected features are processed with K-means clustering method, to distinguish the green tea samples. The classification of green tea samples is performed by using XGBoost (Extreme Gradient Boost) algorithm. The use of XGBoost provides fast converging with limited steps, direct route to the least error and simplified calculations to lower computational costs and improve speed. Finally, the efficacy of the proposed system is calculated based on metrics namely, Calinski Harabasz Score, Distortion score and Silhouette score for K-means clustering technique. The efficiency of the introduced XGBoost method is calculated based on the metrics such as accuracy, AUC (Area under Curve), Recall, Precision, F1-score, Kappa and MCC (Matthews Correlation Coefficient) with outcomes of 0.96, 0.99, 0.96, 0.96, 0.96, 0.94 and 0.95. Thus, the outcome of the proposed method denotes the effective performance in classifying the green tea samples.