关键词: 正交多项式回归, 高蛋白大豆, 施肥, 经济效益, 预测模型

Abstract:

The purpose is to further explore the optimal fertilization pattern combination of high-protein soybean and obtain the best economic benefit. The high-protein soybean ‘Jidou 12’ was used as the test material, and three treatments were set for nitrogen, phosphorus and potassium. With different application amounts of nitrogen, phosphorus and potassium as independent variables and the corresponding soybean economic benefits as dependent variables, the orthogonal polynomial regression method was used to establish the economic benefit prediction model equation of high-protein soybean, and the significance (F) test was carried out for each item. The results showed that potassium and phosphate fertilizer could play a better role in fertilizer efficiency only when nitrogen fertilizer reached a certain level (60-90 kg/hm²). Too low or too high potassium application was not conducive to the increase of soybean economic benefit value. The economic benefit of soybean was positively correlated with the amount of phosphorus applied between 50 and 110 kg/hm². According to F test, nitrogen fertilizer had the largest effect on the economic benefit of high-protein soybean, followed by phosphorus and potassium. The estimated variance was $\hat{σ}$=134.5508, and the 95% confidence interval of economic benefit per hectare of high-protein soybean was predicted to be $\hat{W}$±1.96s=W±263.7196, and the prediction effect was better. Under the model fitting, the application rate of nitrogen, potassium and phosphorus was 54.67, 62.38 and 140 kg/hm², respectively, with the highest soybean economic benefit of 13276.13yuan/hm². The multivariate orthogonal polynomial regression model has practical reference value for predicting the economic benefit of high-protein soybean under multi-factor fertilization.

Key words: orthogonal polynomial regression, high protein soybean, fertilization, economic benefit, prediction model