Abstract 115
Chemometric investigation of barley and malt data
J. Amer. Soc. Brew. Chem. 70 (3): 163-175, 2012
K.J. Siebert, A. Egi and R. McCaig
Several hundred samples of barleys and corresponding pilot scale malts were analyzed for eight barley parameters and 15 malt parameters. Principal components analysis (PCA) was applied to the barley and malt data sets. The barley data had three significant PCs, corresponding to kernel size, germination rate and protein content, and moisture. The malt data had 5 significant components, largely corresponding to modification, extract, enzyme activity, nitrogenous substances, and wort pH. Pattern recognition of the barley and malt data sets was carried out with Linear Discriminant Analysis (LDA), k–Nearest Neighbor analysis (k–NN) and SIMCA. Classification of the barley samples into 2– or 6–row, winter or spring, origin country and cultivar was fairly successful. Classification of the malt samples into hulled or hulless barleys, country of origin, and cultivar was quite successful; classification by crop year and 2– or 6–row barley was less successful. Models of malt parameters as a function of multiple barley measurements were constructed using partial least squares regression (PLSR). An excellent model of malt total protein (R2 = 0.74) was obtained. Fair models of friability, fine and coarse extract, soluble protein, Kolbach index, diastatic power and α–amylase activity were produced. Only poor models of the other parameters were obtained.
Back to K.J. Siebert home page | Back to publication list