Implementation of Principal Component Analysis-Cluster Analysis on The Extraction of Green Tea Leaf (Camellia sinensis (L.) Kuntze)

Abstract

Several green tea leaf extraction studies only describe the results of the extraction method or the different types of solvents that produce the highest levels of polyphenols or caffeine without further analysis by statistical analysis. In addition, the statistical analysis method still often used is a statistical analysis of variance, which has weaknesses. This study used PCA and CA methods to analyze samples based on the solvent's effect on temperature and pH. The solvents used in extracting green tea leaves were hot distilled water at ±95oC, distilled water, citrate buffer pH 4.3, and phosphate buffer pH 7.4 without heating (±25oC). The parameters analyzed were yield, water content, total ash content, acid insoluble ash content, total polyphenol content, and caffeine in green tea leaf extract (Camellia sinensis (L.) Kuntze). Classification with PCA results in a 2-dimensional data reduction that represents all data. Therefore, PC1 can extract as much as 68.7% of the information, and PC2 can extract 22.9% of the information. Cumulatively, PC1 and PC2 extracted as much as 91.5% of the information. Classification with CA resulted in 3 clusters. The third cluster, namely numbers 2 and 3, is the cluster that has the closest similarity with the distance between the cluster centroids of 2.08564 and the similarity level of 56.7850%.