| dc.description.abstract | Quality of medical plants affects the quality of medicine products. To keep the quality and safety of medicines analysis on quality control is needed. This work aims to use the spectral imaging as a method to quality control of sambiloto leaves with various qualities based on age of plant and combined with pattern recognition techniques (PCA and PLSDA). Thin layer chromatography analysis showed different quality of active compounds in sambiloto at age 1, 2, and 3 months. The best standard model and polynomial using Wiener estimation was 97 colours and 46 herbal images and polynomial of order 2. Root mean square error (RMSE) and goodness of fit coefficient values were of 15.21% and 0.8079, respectively. PCA model gave PC1 98% and PC2 1%. PLSDA analysis gave 3 models, namely 1 month (R2 calibration= 0.8640, R2 prediction= 0.8514, RMSE of calibration (RMSEC)= 0.1693, RMSE of prediction (RMSEP)= 0.1821), 2 month (R2 calibration= 0.7570, R2 prediction= 0.7092, RMSEC= 0.2429, RMSEP= 0.2612), and 3 month (R2 calibration= 0.8167, R2 prediction= 0.7958, RMSEC= 0.1929, RMSEP= 0.2070). | en |
