现代电子技术2012，Vol.35Issue(24)：82-84,3.
基于蒙特卡罗无信息变量消除的烟气指标预测
Prediction of tobacco smoke index based on method of Monte Carlo uninformative variables elimination
门月 1丁香乾 2刘孝良1
作者信息
- 1. 中国海洋大学 信息科学与工程学院,山东青岛 266071
- 2. 中国海洋大学 信息工程中心,山东青岛 266071
- 折叠
摘要
Abstract
NIRS analysis method was applied to nondestructive and rapid quantitative analysis of CO, nicotine and tar content of tobacco, because it can improve the prediction precision and eliminate the influence of uninformative variables on model robustness. In this paper, the theory of Monte Carlo uninformative variables elimination ( MC-UVE) is used to select the wavelength variables of near-infrared spectroscopy of tobacco, and then build the partial least squares (PLS) calibration model according to the selected waveband. The results reveal that the modeling variables can be selected by MC-UVE method, which can not only overcome the difficulties that different information intervals of complicated samples have different contribution to PLS model, but also improve the stability of the model and prediction accuracy of multivariate calibration.关键词
近红外光谱/蒙特卡罗无信息变量消除/变量筛选/偏最小二乘法Key words
near infrared spectroscopy/ Monte Carlo uninformative variable elimination/ variable selection/ partial least square分类
信息技术与安全科学引用本文复制引用
门月,丁香乾,刘孝良..基于蒙特卡罗无信息变量消除的烟气指标预测[J].现代电子技术,2012,35(24):82-84,3.
