机构地区: 山东大学数学学院
出 处: 《高师理科学刊》 2007年第3期27-29,共3页
摘 要: 异方差是计量经济工作中线性回归模型经常遇到的问题,异方差的存在对线性回归分析有很强的破坏作用.通过对异方差产生的原因和后果进行分析,利用异方差的戈德菲尔特-夸特检验、拉格朗日乘数(LM)检验、怀特检验方法,判断线性回归模型异方差的存在性.通过加权最小二乘法或可行广义最小二乘法进行修正,建立能够真正反映经济规律的经济模型,实现对经济的正确指导作用. Heteroscedasticity is one of the questions which we meet in measuring analysis frequently. Heteroscedasticity existence has the very strong destructive analysis. Analyzed the reason and the consequence. Through Goldfeld-quandt test, the model of linear regression effect to the linear regression ge multiplier test, white's test, judge heteroscedasticity existence. Carry on the revision through weighted least squares method or feasible generalized least squares method. Thus establishes can truly reflect the economic law and realizes to the economical correct instruction function.
关 键 词: 异方差 戈德菲尔特 夸特检验 拉格朗日乘数 检验 最小二乘法
领 域: [经济管理]