SUN Lei, WU Wenhai, KE Jian, et al. Insulator pollution detection method based on improved SVDD[J]. 2019, 38(9): 57-59. DOI: 10.13873/J.1000-9787(2019)09-0057-03.
In order to solve the problem that while support vector data description is applied in insulator contamination detection
due to edge samples is difficult to distinguish
so false detection and missed detection occur
a data domain description model using double surround hyper sphere is proposed. This method designs a large bounding sphere on the basis of supporting vector data to describe the small bounding sphere. The data are divided into three parts by double bounding sphere: normal area
abnormal area
and rejected area. The rejected area is introduced to give up decisions on the samples which are hard to distinguish. Experimental results show that this method can effectively reduce false detection and missing detection.
Robust support vector data description for outlier detection with noise or uncertain data [J] . Guijun Chen,Xueying Zhang,Zizhong John Wang,Fenglian Li. Knowledge-Based Systems . 2015
Density weighted support vector data description [J] . Myungraee Cha,Jun Seok Kim,Jun-Geol Baek. Expert Systems With Applications . 2014 (7)
Support Vector Data Description [J] . David M.J. Tax,Robert P.W. Duin. Machine Learning . 2004 (1)