|Table of Contents|

Analysis of Generalized Maximum Entropy Regression Effect

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2006年06期
Page:
793-796
Research Field:
Publishing date:
2006-12-30

Info

Title:
Analysis of Generalized Maximum Entropy Regression Effect
Author(s):
HUO Ying-bao~1ZHU Chang-hong~2HAN Zhi-jun~2
1.School of Business and Management,Nanjing University of Finance and Economy,Nanjing 210046,China;2.School of Economy and Management,NUST,Nanjing 210094,China
Keywords:
regression method generalized m ax imum entropy least square m ethod partia l least square method
PACS:
O212
DOI:
-
Abstract:
A im ing at generalizedm ax imum entropy ( GME ) regression effect and especially the indeterm ination o f the cho ice of support space of parameter and erro r in the mode,l the mode lling process o f GME reg ression method is analyzed, and its regression effect is compared w ith others. effects through tw o cases in this paper. Resu lts show that the forecasting prec ision and the exp laining ab ility of GME m ethod is higher than the least square method, the partial least squarem ethod and the princ ipal component ana lysis. The larger the support space of param eter is, the better, under the lack of pr ior in format ion, and the support space o f error should be betw een 3σ and 4σ.

References:

[ 1] M arsh T L, M itte lhamm er R C. Advances in econom etr ics, spatial and spa tiotemporal econom e trics[M ]. New Yo rk: Pace and LeSage, 2004. 203- 238.
[ 2] 霍映宝, 韩之俊. 基于广义最大熵原理和遗传算法 的多指标权重确定方法研究[ J] . 数理统计与管 理, 2005, 24( 3): 39- 44.
[ 3] Go lan A, G zy lH. A genera lizedm axentrop ic invers ion procedure for no isy da ta [ J]. Applied M athema tics and Com putation, 2002, 127( 4) : 249- 260.
[ 4] Lence S H, M iller D J. Estim ation o f mu lt-i output produc tion functions w ith in com plete data: A genera-l ized m ax im um entropy approach [ J]. Eur Rev Ag r Econ, 2002, 25( 12): 188- 209.
[ 5] CaputoM R, Par is Q. Com pa ra tive statics of the genera lized m ax imum entropy estim ator of the genera l linear model[ Z] . Dav is: Departm ent o f Agricultura l and Resource E conom ics Un ive rs ity of Californ ia) Dav is, 2000.
[ 6] 王惠文. 偏最小二乘回归方法及其应用[M ]. 北 京: 国防工业出版社, 1999.
[ 7] 霍映宝, 韩之俊. 三种有偏回归方法的比较分析 [ J]. 统计与决策, 2002, 188( 11): 9- 10.

Memo

Memo:
-
Last Update: 2006-12-30