【摘要】 应用偏最小二乘回归方法,构建了样本数小于自变量数的复杂土地利用系统变化的驱动机制模型。沙坪坝区的研究结果表明:总人口、国内生产总值和全社会固定资产投资额是推动该区土地利用系统变化的最主要驱动力,非农人口和人均国内生产总值是次强驱动力,农林牧渔产值对整个土地利用系统变化的驱动作用最弱。基于回归原理和回归模型,全面比较了逐步回归和偏最小二乘回归的特点,逐步回归自变量存在取舍,不能全面反映各驱动因子对系统变化的作用,偏最小二乘回归具有结论更加可靠、整体性更强的特点,更利于实际应用。更多还原

【Abstract】 A dynamics model of quite integrated land use change is established through partial least squares regression,in which sample number is less than independent variables.Comprehensive analyses indicate that population,gross domestic product and total investment of fixed asset are the major drive power of land use change in Shapingba District;urban population and gross domestic product per capita are the intermediate drive power;output value of agriculture,forestry,husbandry and fishery are the minor drive power of land use change.Based on regression principle and models,stepwise regression and partial least squares regression are compared.The result indicates that independent variables are selective by stepwise regression,and therefore cannot reflect each drive factor′s effect upon system.Partial least square regression is more reliable and systematic,thus it is widely applied.更多还原