Multivariate Analysis of Mash Data

Different mash plant traits contributes to the mash grain yield but the major contributors are plant height (X1), days to flowering (X2), days to first pod maturity (X3), days to 90% maturity (X4), branches per plant (X5), pods per plant (X6), pod length (X7), seeds per pod (X8), 100-seed weight (X9), biological yield per plant (X10) and mash grain yield (Y). This study was initiated to find the important regressors on which the yield of mash depends. In this regard principal component analysis and path analysis were used to find correlation structure between mash plant traits and regressors effect on mash grain yield, respectively. Principal component analysis reduced the dimensionality in the system of eleven mash plant traits to four principal components, which contributes about 88% of the total variability present in the mash plant data. On the basis of correlation between principal components and original mash plant traits, a classification structure was made to observe the relation between different traits. It was observed that for the first principal component, plant height (X1), days to flowering days to first pod maturity (X3), days to 90% maturity (X4) and 100 seeds weight (X9) have positive correlation between themselves i.e. varies in the same direction. Path analysis was also described to explain correlation structure, direct-indirect effects between different mash plant traits. This analysis suggested that pod per plant has maximum positive direct effect on mash grain yield i.e. more pod per plant, greater will be the yield. But days to 90% maturity has maximum negative direct effect on mash grain yield i.e. more maturity lesser will be the grain yield. Similarly branches per plant and biological yield per plant have positive indirect effect on mash grain yield via pods per plant. It was observed that the direct and indirect effects of remaining predictors are negligible.