|
|||||
|
|
| 融合互信息与线性变换的非线性特征提取 | |
| 引用本文： | 徐洪章,郝宁波,廖海斌. 融合互信息与线性变换的非线性特征提取[J]. 计算机工程与应用, 2011, 47(36): 222-225. DOI: 10.3778/j.issn.1002-8331.2011.36.061 |
| 作者姓名： | 徐洪章 郝宁波 廖海斌 |
| 作者单位： | 1. 黄淮学院计算机科学系,河南驻马店,463000 2. 黄淮学院国际学院,河南驻马店,463000 3. 武汉大学电子信息学院,武汉,430079 |
| 基金项目： | 国家自然科学基金(the National Natural Science Foundation of China under Grant No.40930532); 郑州市重大科技攻关项目(No.072SGZS38042) |
| 摘 要： | 由于线性变换无法较好保留数据的非线性结构而非线性变换往往需要进行大量的复杂运算,提出一种快速、高效的非线性特征提取方法。该方法通过研究互信息梯度在核空间中的线性不变性,采用互信息二次熵快速算法及梯度上升寻优策略,在有效降低计算量的同时能够提取有判别力的非线性高阶统计量。详细的数据投影和分类实验表明该方法在分类性能和算法时间复杂度上都优于传统算法。 |
| 关 键 词： | 核方法 非线性变换 特征提取 互信息 |
| 修稿时间： | |
Nonlinear feature extraction by fusing mutual information and linear transformation |
|
| XU Hongzhang,HAO Ningboz,LIAO Haibin. Nonlinear feature extraction by fusing mutual information and linear transformation[J]. Computer Engineering and Applications, 2011, 47(36): 222-225. DOI: 10.3778/j.issn.1002-8331.2011.36.061 | |
| Authors: | XU Hongzhang HAO Ningboz LIAO Haibin |
| Affiliation: | XU Hongzhang,HAO Ningbo,LIAO Haibin.1.Department of Computer Science,Huanghuai University,Zhumadian,Henan 463000,China 2.School of International College,China 3.School of Electronic Information,Wuhan University,Wuhan 430079,China |
| Abstract: | Linear transformation can not better retain the nonlinear structure of data,but the nonlinear transformation often requires lots of complex measurements.To address this,a fast and effective method of nonlinear feature extraction is proposed. This method studies the linear invariance of mutual information gradient in the kernel space,and employs a fast algorithm for mutual information and gradient ascent.In this way,the extracted features can reflect the characteristics of discriminative higher-order statist... |
| Keywords: | kernel method nonlinear transformation feature extraction mutual information |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |
| 点击此处可从《计算机工程与应用》浏览原始摘要信息 | |
| 点击此处可从《计算机工程与应用》下载免费的PDF全文 | |