Classification of agarwood oil quality using random forest and grid search crossvalidation / Mohamad Aqib Haqmi Abas ...[et al.]

This paper presents a machine learning technique to classify the agarwood oil quality. The random forest classifier model is used with the grid search cross validation technique to classify the quality of agarwood oil. The data of agarwood oil sample were obtained from Forest Research Institute Malaysia (FRIM) and Universiti Malaysia Pahang, Malaysia. In this experiment, the chemical compound abundances information of the agarwood oil that has been extracted from GC-MS machine is used as the input feature and the quality of the sample oil which is high quality and low quality is used as the output feature. Based on the result obtained from this study, using Gini impurity measure as criterion combined with 3 level maximum depth of decision trees and 3 number of maximum features for each tree provides the best classification accuracy of the agarwood oil quality sample at 100% and performance measure scores of 1.0.