Correlation analysis and modeling application from objective indicators to subjective evaluation of scented tea: A case study of rose tea.

Different scented teas provide various choices for consumers from appearance, aroma, flavor and others. Aiming to define advantages and market positions of different scented teas and promote optimization of market structure, characteristics for scented tea favored by consumers and outstanding attributes of different scented teas should be clarified. Rose tea was taken as study object. Sensory evaluation and consumer acceptance were investigated. GC-MS and HPLC fingerprints were established. Physicochemical characteristics were determined. RGB integration analysis was inventively proposed for correlation analysis. The volatile compounds with spicy, green or herbal odor as camphene, β-phenethyl acetate, eugenol, and physicochemical parameters as antioxidant capacity, reducing sugar content, pH showed positive correlation with popular sensory properties. Six models for consumer preference by objective description were built through GA-SVR (accuracy = 1), and APP was developed. The research mode of scented tea has been successfully established to study multiple subjective characteristics with measurable objective parameters.