Array
(
[session_started] => 1733714300
[LANGUAGE] => EN
[LEPTON_SESSION] => 1
)
A STATISTICAL APPROACH FOR ESTIMATING WHEAT YIELD USING BOOTSTRAP RESAMPLING FOR RAIN-FED FARMING: A CASE STUDY OF KURDISTAN PROVINCE, IRAN
S. Bazgeer, A. Khoorani, P. Zeaeian, M. Farajzadeh
Abstract: For the purpose of modeling and predicting rainfed wheat (Triticum aestivum) yield in Kurdistan province, Iran, five weather parameters, as well as three agrometeorological indices were used, as independent variables in linear regression models during 1991-2003. The independent variables were extracted for different phenological phases during the plant-growing season from sowing to harvest. Backward regression models were used to model rain-fed wheat yield and sensitivity analysis was carried out on the models. on the basis of choosing the best models for each district and Kurdistan province (in the north west of Iran), the bootstrap resampling method was run on them. Both above-mentioned models were validated for 2003-2006 years data by estimating the rain-fed wheat yield. The results show that using bootstrap resampling method for modeling and
estimating the crop yield increases the interior accuracy (increasing r, multiple correlation coefficient, from 0.84 to 0.98, and decreasing SEOE, standard error of estimate, from 166 to 47 kg/ha) of the models.
Keywords: bootstrap; rain-fed wheat; regression models; yield estimation
Date published: 2019-04-18
Download full text