|
|||||
|
|
| 基于小波分析和非线性映照的多维异常样本检测方法 | |
| 引用本文： | 宋彦坡,唐英,彭小奇. 基于小波分析和非线性映照的多维异常样本检测方法[J]. 系统仿真学报, 2006, 18(4): 978-981 |
| 作者姓名： | 宋彦坡 唐英 彭小奇 |
| 作者单位： | 1. 中南大学信息科学与工程学院,长沙,410083;中南大学能源科学与工程学院,长沙,410083 2. 中南大学物理科学与技术学院,长沙,410083 |
| 基金项目： | 中国科学院资助项目;高等学校博士学科点专项科研项目 |
| 摘 要： | 数据预处理对于数据挖掘及其它基于数据样本的系统建模极为重要。数据预处理的一项重要任务是从大量数据样本中剔除异常样本,但在数据集中各数据项间关系未知的情况下,检测异常样本比较困难,为此,提出了一种基于小波分析和非线性映照的异常数据样本检测方法。该方法通过非线性映照对多维数据进行降维处理后,利用小波分析的局部分析优势检测异常样本。仿真结果表明,该方法切实可行,效果良好,有较强的实用性。 |
| 关 键 词： | 数据预处理 非线性映照 小波分析 异常检测 数据挖掘 |
| 文章编号： | 1004-731X（2006）04-0978-04 |
| 收稿时间： | 2005-02-22 |
| 修稿时间： | 2005-09-23 |
Approach Based on Wavelet Analysis and Non-linear Mapping to Detect Anomalies in Dataset |
|
| SONG Yan-po,TANG Ying,PENG Xiao-qi. Approach Based on Wavelet Analysis and Non-linear Mapping to Detect Anomalies in Dataset[J]. Journal of System Simulation, 2006, 18(4): 978-981 | |
| Authors: | SONG Yan-po TANG Ying PENG Xiao-qi |
| Affiliation: | 1. School of Information Science and Engineering, Central South University, Changsha 410083, China; 2. School of Energy Science and Engineering, Central South University, Changsha 410083, China; 3. School of Physics Science and Technology, Central South University, Changsha 410083, China |
| Abstract: | Data preprocessing is crucial for data mining or other system modeling based on data set.It is an important task to eliminate the anomalous samples which have been polluted.But it's difficult to detect the anomalies before the relationship among the samples' attributes is known.Therefore,an approach based on wavelet analysis and non-linear mapping to detect anomalies was proposed.Using the non-linear mapping to decrease the dimensions of data,taking full advantage of wavelet analysis' superiority in local analysis,the approach is able to detect anomalies accurately.The experiments show that the approach is accurate and practical. |
| Keywords: | data preprocessing non-linear mapping wavelet analysis anomaly detecting data mining |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ | |