基于偏最小二乘回归的农村居民人均纯收入预测.pdfVIP

27 2 Vol1 27, No1 2 2011 5 SYSTEM SC IENCES AND COMPREHENSIVE STUDIES IN AGRICULTURE M ay, 2011 , (东北农业大学 工程学院, 黑龙江 哈尔滨 1500 0) : 偏最小二乘回归是一种能在 一个算法下 时实现回归建模数据简化和两组变量间相关分析的新型多元回归分析 方法农村居民人均纯收入受到多种因素的影响, 各因素间相互制约和影响, 存在着严重的多重相关性, 而偏最小二乘回 归方法能很好地解决这 一问题通过将偏最小二乘回归分析方法用于农村居民人均纯收入拟合和预测时, 结果表明, 该方 法具有建模简单计算结果可靠的特点, 具有较强的实用性图 2, 表 4, 参 6 : 偏最小二乘回归; 成分提取; 人均纯收入; 预测 + : S11 4 : A : 1001- 0068 ( 2011) 02- 0142- 06 Rural Per Capita Net Incom e ForecastBased on Partial Least Squares Regression SUO Ru-ixia, WANG Fu-lin (Engineering College, N ortheastA griculture University, H arbin 1500 0, China) Abstract: Partial least squares regression is a new mu ltiple regression analysis m ethod that can achieve smi ultaneously in a regress ion m odeling, data reduction and correlation analysis betw een wt o sets of variables1 Ru ral per cap ita net incom e is affected by many fac- tors, w hich the factors constraint and affect each other, and there is a serious mu ltiple correlation, and partial least squares regress ion m ethod ( PLS) can w ell solve the problem 1 The PLS is used for fitting and pred iction of rural residents per cap ita net incom e, the re- su lts show that the PLS process ofm odeling is smi p le and the precision of pred iction is reliable, w ith strong practicability1 K ey words: partial least squares regress ion; com ponent extraction; per cap ita net income; forecast 0 引 言 [ 1] , , / 0 2009 2 1, 56, , , , , , S1W old [ 2] , , , , , , [ - 4] 1 偏最小二乘回归模型 111 : , : 2010 - 06- 0 ; : 2010- 07