摘要
支持向量数据描述(Support Vector Data Description,SVDD)被认为是用于异常检测的典型方法。众所周之,参数的设置和特征的品质是影响SVDD性能的两个关键点。将SVDD的特征提取和参数选择问题结合在一起,提出了一种基于模拟退火的SVDD特征提取和参数选择方法(SA-SVDD)。在模拟退火的过程中,自动选择最优核参数、折衷参数以及抽取特征的维数。在UCI基准数据集上的实验结果表明,与传统的参数选择方法相比,SA-SVDD取得了更优的性能。
Support vector data description(SVDD) is considered as a classical method for novelty detection.As is well known,the parameter setting and the quality of features are two key points to affect the performance of SVDD.Combining feature extraction and parameter selection of SVDD,this paper proposed a simulated annealing approach for feature extraction and parameter selection of SVDD(SA-SVDD).During the procedure of simulated annealing,the optimal kernel parameter,trade-off parameters,and number of extracted features are automatically selected.Experimental results on the UCI benchmark data sets demonstrate that SA-SVDD has better performance than the traditional parameter selection methods.
出处
《计算机科学》
CSCD
北大核心
2013年第1期302-305,共4页
Computer Science
基金
国家自然科学基金项目(60903089
61073121
61170040)
河北大学基金项目(2008123
3504020)资助
关键词
特征提取
模拟退火
参数选择
SVDD
异常检测
Feature extraction
Simulated annealing
Parameter selection
SVDD
Novelty detection