STUDIES ON THE USE OF UV-VIS SPECTROSCOPY FOR IDENTIFICATIONOF BLENDING OF CIVET COFFEE WITH ARABICA COFFEE

This study aims to identify the authentication of civet coffee using a Soft independent modeling of class analogy (SIMCA) method and principal component analysis (PCA). The test carried out on the coffee powder measuring 0.297 millimeters (mesh 50). Comparison of blend that is samples 1- 50 each 1 g of pure civet coffee, samples 51- 60 each 0.9 g civet coffee and 0.1 g arabica coffee, samples 61-70 each 0.8 g civet coffee and 0.2 g arabica coffee, samples 71-80 each 0.7 g civet coffee and 0.3 g arabica coffee, samples 81-90 each 0.6 g civet coffee and 0.4 g arabica coffee, samples 90-100 each 0.5 g civet coffee and 0.5 g arabica coffee. The classification results show SIMCA and PCA methods are able to identify civet coffee mixture. PC 1 explains 75% the variance of data and PC2 explains 17% the variance of data. Values obtained on SIMCA classification are specificity 76%, sensitivity of 84% and accuracy of 80%, with a value error of 23%.