Advanced regression analysis to mitigate multi-collinearity among yield influencing factors under Stemphylium blight stress in Lens culinaris
Pavithra, S.
Advanced regression analysis to mitigate multi-collinearity among yield influencing factors under Stemphylium blight stress in Lens culinaris - Dordrecht (Netherlands) : Springer, 2024.
Peer review
The upsurge of stemphylium blight disease noticed during recent cropping years is the prime global threat for lentil (Lens culinaris Medik.) production. Identification of factors that influence lentil yield with the help of an advanced regression model will speed up the progress of lentil crop improvement for biotic stress tolerance. In this context, an experiment was undertaken to identify the key control factors of lentil yield under stemphylium blight stress. The field experiment was laid out under alpha lattice design using fifty lentil genotypes with two replications. An advanced dimension reduction cum regression approach Partial Least Square Regression (PLSR) was employed to mitigate the effect of multi-collinearity among 23 yield-influencing traits along with traditional Stepwise Multiple Linear Regression (SMLR). The results of SMLR analysis indicated that pods per plant, number of seeds per pod, hundred seed weight, superoxide dismutase and pod yield per plant had considerable effects on seed yield per plant with the R-squared value of 0.940. The first four PLSR components were considered to be optimum which were cumulatively explained 93.10% of the total variance towards lentil seed yield. The trait pods per plant was recorded with the highest PLSR regression coefficient devoid of multi-collinearity effects among the independent yield attributing variables under stemphylium blight environment and hence concluded to be the most influencing trait towards lentil seed yield followed by seeds per pod, hundred seed weight, pod yield per plant and superoxide dismutase.
Text in English
0014-2336 1573-5060 (Online)
https://doi.org/10.1007/s10681-024-03382-7
Lentils
Multiple regression analysis
Stemphylium
Blight
Advanced regression analysis to mitigate multi-collinearity among yield influencing factors under Stemphylium blight stress in Lens culinaris - Dordrecht (Netherlands) : Springer, 2024.
Peer review
The upsurge of stemphylium blight disease noticed during recent cropping years is the prime global threat for lentil (Lens culinaris Medik.) production. Identification of factors that influence lentil yield with the help of an advanced regression model will speed up the progress of lentil crop improvement for biotic stress tolerance. In this context, an experiment was undertaken to identify the key control factors of lentil yield under stemphylium blight stress. The field experiment was laid out under alpha lattice design using fifty lentil genotypes with two replications. An advanced dimension reduction cum regression approach Partial Least Square Regression (PLSR) was employed to mitigate the effect of multi-collinearity among 23 yield-influencing traits along with traditional Stepwise Multiple Linear Regression (SMLR). The results of SMLR analysis indicated that pods per plant, number of seeds per pod, hundred seed weight, superoxide dismutase and pod yield per plant had considerable effects on seed yield per plant with the R-squared value of 0.940. The first four PLSR components were considered to be optimum which were cumulatively explained 93.10% of the total variance towards lentil seed yield. The trait pods per plant was recorded with the highest PLSR regression coefficient devoid of multi-collinearity effects among the independent yield attributing variables under stemphylium blight environment and hence concluded to be the most influencing trait towards lentil seed yield followed by seeds per pod, hundred seed weight, pod yield per plant and superoxide dismutase.
Text in English
0014-2336 1573-5060 (Online)
https://doi.org/10.1007/s10681-024-03382-7
Lentils
Multiple regression analysis
Stemphylium
Blight