Abstract

Hinoki hydrosol is the remaining by-product of the distillation process of hinoki essential oil. In traditional applications, although hinoki hydrosol has many health benefits, due to the different concentrations and uncontrollable changes, most of the hinoki hydrosol is discarded and not used, in order to achieve the purpose of circular economy utilization. In order to distinguish the concentrations of hinoki hydrosol and pure water samples, this study designed a set of experimental methods to scan the spectra of various hinoki hydrosol and pure water samples, pre-process the spectra, and then apply two types of machine learning Calculation methods (K-means and PLSR).

Hinoki hydrosol is the remaining by-product of the distillation process of hinoki essential oil. In traditional applications, although hinoki hydrosol has many health benefits, due to the different concentrations and uncontrollable changes, most of the hinoki hydrosol is discarded and not used, in order to achieve the purpose of circular economy utilization. In order to distinguish the concentrations of hinoki hydrosol and pure water samples, this study designed a set of experimental methods to scan the spectra of various hinoki hydrosol and pure water samples, pre-process the spectra, and then apply two types of machine learning Calculation methods (K-means and PLSR).

In the future, follow-up research will plan to combine GC-MS experimental instruments and obtain more accurate concentration data to train the PLSR model studied, so that the improved PLSR model can further improve the accuracy of quantitative prediction of hinoki hydrosol. And conduct research on the subsequent application of hinoki hydrosol products.

Keywords:Hinoki hydrolat; K-means; GC-MS; Spectra; K-means; PLSR