论文部分内容阅读
应用“3414”设计的田间肥料试验结果,探讨蔬菜氮磷钾三元肥效模型的Monte Carlo建模法。莴苣试验结果表明,采用最小二乘法回归建模,NP、NK二元肥效模型属于非典型式,PK二元和NPK三元肥效模型属于典型式;而改用MonteCarlo建模法,NP、NK和PK二元肥效模型以及NPK三元肥效模型均属于典型式。采用最小二乘法回归建模,不同蔬菜作物建立的34个三元肥效模型的典型式出现机率为41.2%,而Monte Carlo建模法的典型式出现机率为91.2%,提高了2.2倍。Monte Carlo建模法是适当放弃数学上偏差平方和最小的最优性,使待估参数达到专业上最优而数学上较优,从而提高典型肥效模型的出现机率。对莴苣氮钾非典型肥效模型的推荐施肥表明,Monte Carlo法的结果明显优于产量频率分析法。因此,Monte Carlo建模法为建立蔬菜多元肥效模型和推荐施肥提供了一种有效方法。
Application of “3414” design of field fertilizer test results to explore the vegetable nitrogen, phosphorus and potassium ternary efficiency model of Monte Carlo modeling. The results of lettuce experiment showed that the NP and NK binary efficiency models belonged to the atypical form using least-squares regression modeling, and the PK binary and NPK ternary models belonged to the canonical model; instead, Monte Carlo modeling, NP, NK and PK The dual fertilizer efficiency model and the NPK three-factor fertilizer efficiency model belong to the typical model. Using least-squares regression modeling, the typical occurrence probability of 34 ternary efficiency models established by different vegetable crops was 41.2%, while the typical occurrence probability of Monte Carlo modeling method was 91.2%, an increase of 2.2 times. The Monte Carlo modeling method is to give up the minimum optimality of the sum of squares of deviations reasonably, so that the parameters to be estimated are optimally and mathematically better, so as to improve the probability of emergence of a typical fertilizer model. The recommended fertilization for the atypical nitrogen and potassium fertilizer efficiency model of lettuce showed that the result of Monte Carlo method was obviously better than the yield frequency analysis method. Therefore, Monte Carlo modeling method provides an effective method for establishing multi-fertilizer efficiency model of vegetables and recommending fertilization.