Use of multivariate analysis and machine learning methods to characterize traits contributing to wheat yield diversity

Aim of study: Regarding the third largest staple food crop in world, determining factors affecting wheat yield is great importance. This study aimed to determine useful subsets agronomic traits and evaluate order importance grain yield.
 Area Fars province, Iran.
 Material methods: In total, data corresponding 22 was collected from six different regions (Darab, Kavar, Marvdasht, Fasa, Lar, Khonj) 90 farms Iran as most important wheat-growing regions. Multivariate statistical analysis (correlation, stepwise regression, principal component (PCA)) machine learning modeling approaches, such partial least squares regression (PLSR) support vector (SVR) models, were applied traits.
 Main results: The findings, based on integrated approaches correlation, PCA, highlighted that number spikes m-2, spike-1, thousand-grain weight had a major impact followed by awn length, spike narrow leaf herbicide, broadleaf time plant maturity (month), soil salinity. Besides, PLSR with nine inputs (nine selected traits) displayed better prediction capability (R2=85 %, RMSE=0.32, MSE=0.10, BIAS=-0.05) than all twenty-two input Research highlights: Integrated multivariate analyses methods could be powerful tool have significant yield. These achievements can considered for future breeding programs.