Polynomial tuned Kernel Parameter in SVM of Agarwood Oil for Quality Classification

As part of ongoing research for agarwood oil quality classification, this paper presents the non-linear SVM modelling with polynomial as kernel parameter. The work involves of 96 agarwood oil collection, from different high qualities. The input for SVM modeling is the abundances (%) of volatile and the output is agarwood oil qualities either low or high. The experimental works are carried out automatically via MATLAB software version R2016a. The result showed that polynomial tuning kernel parameter has its capability in classifying agarwood oil volatile to high and low qualities. It was supported by 100 % obtained for accuracy, confusion matrix, sensitivity, precision and specificity. The finding in this study is important and benefits other future work focusing on agarwood oil research area.