PENALIZED SELECTION OF VARIABLE CONTRIBUTING TO ENHANCED SEED YIELD IN MUNGBEAN (Vigna radiata L.)

Penalized regression methods for simultaneous variable selection and coefficient estimation have received a great deal of attention in recent years. Especially those based on the least absolute shrinkage and selection operator (LASSO), that involves penalizing the absolute size of the regression coefficients. The ordinary least square and LASSO methods were used for selection of most significant traits contributing towards seed yield in mungbean plants with 18 morphological and yield associated traits and to develop the prediction model . Bayesian information criterion was applied to choose minimum tuning parameter. Results indicated that dry weight biomass and harvest index were highly significant characters towards seed yield while days to maturity, days to flowering, number of nodes per plant, pods per plant and degree of indetermination had a significant affect on response variable. Based on the results, it was rational to conclude that high yield of mungbean crop could be obtained by selecting the breading materials with these important characters on seed yield.