Application of Supervised Feature Selection Methods to Define the Most Important Feature on Sink/Source Relationships in Maize

Kernel number per unit land area is the most important yield component in maize kernel weight is also an important contributor to kernel yield. Final kernel weight is closely related to the maximum kernel water content (MKWC) achieved during grain filling. This study presents the results of applying supervised feature selection algorithms in the selection of the most important traits contributing to MKWC as a major yield component. Data were obtained from a field experiment conducted during growing season, at the Experimental Farm of the College of Agriculture, Shiraz University, Badjgah, the experimental design was randomized completeblocks with three replicates and the treatments in a split-split plot arrangement, and from literature. Experiments on the subject of sink/source relationships in maize from twelve fields (as records) from the different parts of the world which were different in 22characteristics (features) The feature selection algorithm demonstrated that14 features including: planting date (days), countries, hybrids, P applied (kg/ha), final kernel weight (mg), soil type, season duration (days), days to silking, leaf dry weight (g/plant), mean kernel weight (mg), cob dry weight (g/plant), kernel number per ear, N applied (kg/ha), and duration of the grain filling period ( C day) were the most effective traits in determining maximum kernel water content. Among the effective traits (features), planting date (days) was the most important one. Our results showed that features classification by supervised feature selection algorithms can be used to clarify the important traits contributing to maize kernel water content and yield providing a comprehensive view. This study identified the future research areas in feature selection in maize physiology, introduces newcomers to this field.

Keywords:

Supervised feature selection algorithms , Maximum kernel water content , yield