Abstract
Prediction fat of coffee beans in West Sumatra is still done manually by chemical methods, time-consuming, and costly. With the development of food technology, the estimation of the content of the beans to do with how quickly and efficiently using NIRS technology. NIRS can analyze the chemical content of foodstuffs nondestructive. This study uses Arabica beans Solok Radjo two maturity levels are 4 (ripe) and 5 (over ripe), 30 samples with weight of samples 6 grams. NIRS instrument used is the science and technology FT-IR-1516 with a wavelength of 1000-2500 nm. Fat tests conducted in the laboratory Instrumentation Technology Center for Agriculture. In this study, the data spectral coffee beans processed using PCA (Principal Component Analysis) after correlated with fat content processed by the method of PLS (Partial Least Square). The prediction results of calibration best fat obtained r = 0.999, R2 = 0.999, SEC = 0.008%, and RMSEC = 0.008%, validating the value of r = 0.998, R2 = 0.999, SEP = 0.009%, and RMSEP = 0.008%, of the model fat coffee beans can be identified by non-destructive evaluation using spectral analysis.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have