[1]YAN Xue-feng.Modified nonlinear ridge regression with optimal ridge pa rameter and its application to 4-CBA soft sensor[J].CAAI Transactions on Intelligent Systems,2006,1(1):74-78.
Copy

Modified nonlinear ridge regression with optimal ridge pa rameter and its application to 4-CBA soft sensor

CAAI Transactions on Intelligent Systems[ISSN 1673-4785/CN 23-1538/TP] Volume: 1 Number of periods: 2006 1 Page number: 74-78 Column: 学术论文—人工智能基础 Public date: 2006-03-25
Title:
Modified nonlinear ridge regression with optimal ridge pa rameter and its application to 4-CBA soft sensor
Author(s):
YAN Xue-feng
Automation Institute, East China University of Science and Technology, Shanghai 200237, China
Keywords:
ridge regression;  ridge parameter;  evolution algorithm;  soft sensor
CLC:
O212.4
DOI:
-
Abstract:
Considering that there exist many factors having high -nonlinear and co mplex effect on the concentration of the 4carboxybenzaldehyde (4-CBA) in prod uc t, which was the most important intermediate product of p-xylene oxidation reac ti on, a modified nonlinear ridge regression (MNRR) based on optimal ridge paramet er, was proposed to develop the 4-CBA concentration soft sensor. Satisfactory r e sults were obtained. Firstly, MNRR applied the nonlinear transformation for init ial pattern independent variables to expand pattern space. Secondly, considering that there exists correlation or multicollinearity among the variables in the e xpanding pattern space, the ridge regression was employed, in which evolution al gorithm was used to obtain the global optimal ridge parameter according to the p re dicting ability of the model. Thus, the model was obtained that can describe com plex nonlinear system and has good predict accuracy. The comparison results show that the MNRR model has better predict accuracy than nonlinear least square r egr ession and nonlinear ridge regression based on generalized cross-validation of s electing the ridge parameter. In addition, MNRR overcomes the main flaw in tradi tional ridge regression that is difficult to obtain the global optimal ridge pa rameter, and thus has the robust character.
References:
［1］YAN X F, DU W L,QIAN F. Development of a kinetic model fo r industrial oxidation of pXylene by RBF-PLS and CGA［J］. AIChE Journal, 200 4, 50(6): 1169-1176.
［2］PRENGLE J H W, BAROND N. Make petrochemicals by liquid phase oxidation［J］. Hydrocarbon Process, 1970, 49(3): 106-118.
［3］RAGHAVENDRACHAR P, RAMACHANDRAN S. Liquidphase catalytic oxidation of p-Xylene ［J］. Ind Eng Chem Res, 1992, 31(2): 453-462.
［4］颜学峰, 余?? 娟, 钱??? 锋.优选优生进化算法及4-CBA软测量模型参数估计［J ］. 高校化学工程学报, 2005, 19(2): 238-243.
YAN Xuefeng， YU Juan，QIAN Feng. An evolution algorithm with selectbest an d prepotency operator and parameter estimation of 4-CBA model ［J］. Journal of Chemical Engineering of Chinese Universities, 2005, 19(2): 238-243.
［5］何永勇, 褚福磊, 种秉林. 基于进化计算的神经网络设计与实现［J］. 控制与决策, 2001, 16(3): 257-262.
HE Yongyong，CHU Fulei，ZHONG Binglin. Artificial neural networks design a nd implementation based on evolutionary computation ［J］. Control and Decision, 2001, 16(3): 257-262.
［6］赵启林, 陈??? 斌, 卓家寿. 前馈神经网络结构自删除算法的研究［J］. 河海大学学报, 2000, 28(4): 63-66.
ZHAO Qilin，CHEN Bin，ZHUO Jiashou. Selfdeleting algorithm of feedforwa rd neural networks ［J］. Journal of Hehai University, 2000, 28(4): 63-66.
［7］杨????楠. 岭回归分析在解决多重共线性问题中的独特作用［J］. 统计与决策, 2004(3)：14-15.
?YANG Nan. Role of ridge regression to solve multilinear problem ［J］. St atistics and Decision, 2004(3): 14-15.
［8］GAVIN C， CAWLEYA, NICOLA L C T, ROBERT J F, et al. Heter oscedastic kernel ridge regression ［J］. Neurocomputing, 2004, 57(1): 105-124. 
［9］YAN X F, YU J, QIAN F. The feature-preserving map of high-dimensi onal complex chemical objects using nonlinear map Integrated with correlative c omponent analysis ［J］. Chemometrics and Intelligent Laboratory Systems, 2005, 75(1): 13-22.
［10］孙静民. 聚酯工艺［M］. 北京: 化学工业出版社, 1985.
?SUN Jingmin. Ployester technics ［M］. Beijing: Chemical Industry Press, 1985.
［11］胡上序, 陈德钊. 观测数据的分析与处理［M］. 杭州: 浙江大学出版社, 19 96.
?HU Shangxu, CHEN Dezhao. Observed data analysis and processing［ M］. Hangzhou: Zhejiang University Press, 1996.
［12］MICHALEWICZ Z. Genetic algorithms+data structures=evolution program s［M］. Berlin: Springer-Verlag, 1992.
Similar References:

Memo

-

Last Update: 2009-04-07

Copyright © CAAI Transactions on Intelligent Systems