地理权重回归在土壤pH空间预测中的应用

为高效利用土壤和开展精准农业生产，采用地理权重回归方法对区域土壤pH的空间分布进行预测，以获得比普通线性回归方法更高的预测精度和论证其对其他土壤属性进行预测的可行性。实测样点土壤的pH，然后选用与土壤pH相关程度较高且易于获取的环境因子(海拔高程、坡度、归一化植被指数、铁氧化物指数、样点距河流的最近距离、土壤侵蚀强度)来建立与土壤pH的回归方程。因子经过逐步回归及多重共线性检验后，用地理权重回归模型对pH进行预测。通过ME、RMSE及回归决定系数等指标，用验证样点数据对预测效果进行评价。结果表明，地理权重回归模型的预测精度高于普通线性回归模型，能大幅度降低AIC值，显著提高模型的回归决定系数，有效地降低回归的残差值；生成的预测图在空间上过渡自然，制图效果较好；利用地理权重回归法能将环境影响因子纳入模型，并能充分考虑样点的空间位置，适合于具有高度空间变异的土壤属性预测及制图。

Application of geographically weighted regression onthe spatial prediction of soil pH

Based on data of 273 Chinese enterprises, this study conducted an evaluation index system of harmonious labor relations consisting of 5 first-level indicators and 25 second-level indicators, and set criteria in identifying levels of enterprise harmonious labor relations. The evaluation index system was applied in nine enterprises labeled as national model of labor relations. Results showed that labor relations were most harmonious in social security and least harmonious in wage distribution, and were moderate harmonious in training and working environment, evaluation of labor contracts, union construction and labor dispute resolution. Therefore labor relations of Chinese enterprises were still in need of further adjustment and improvement.