STATISTICAL-ANALYSIS OF TEA SENSORY DATA - METHOD AND APPLICATION

As the purpose of this paper, we are going to analyze the human sensor y evaluation data by using balance and unbalance data analysis of vari ance (ANOVA) methods, which are rarely used to aware the influential f actors of quality, directly. Not only use F and approximate F (F') tes t to identify the character of each factors, but also apply Duncan mul tiple comparison method to determine the difference among means of the factor. These statistical methods will give us an objective result of human sensory data, put us in reference to quality evaluation, develo pment of high quality products, and cultivar selection. In this paper, we are going to compare the quality of tea among several kinds of var ieties which are cultivated by Taiwan Tea Experiment Station (TTES). W e use three methods to know the effect of influential factors of tea q uality. The first method is analyzed by factors of seasons and varieti es at each counties in each years, the second method is analyzed by di fferent types of tea in each years, the third method is analyzed by di fferent types among these years. Depending on the analysis of the firs t method, it gives us different results in different counties. It is h ard to get a conclusion of which cultivar is the best, because the qua lity is affected by seasons, temperature, environment, disease, making process and so on in different planted area. Under the analysis of th e second and the third methods, TTES12 and Ching-shin Oolong have the best quality of tea among all varieties, TTES 14 and TTES 16 are the s econd and TTES 15 and TTES 17 are the third