A Spectral Clustering Algorithm for Outlier Detection

Recently, spectral clustering has become one of the most popular modern clustering algorithms which are mainly applied to image segmentation. In this paper, we propose a new spectral clustering algorithm and attempt to use it for outlier detection in dataset. Our algorithm takes the number of neighborhoods shared by the objects as the similarity measure to construct a spectral graph. It can help to isolate outliers as well as construct a sparse matrix. We compare the performance of our algorithm with the k-means based clustering algorithm while using them to detect outliers. Experiment results show that the algorithm can obtain stable clusters and is efficient for identifying outliers.

[1]  Francesco Masulli,et al.  A survey of kernel and spectral methods for clustering , 2008, Pattern Recognit..

[2]  Jianbo Shi,et al.  Learning Segmentation by Random Walks , 2000, NIPS.

[3]  Marina Meila,et al.  A Comparison of Spectral Clustering Algorithms , 2003 .

[4]  Shaogang Gong,et al.  Spectral clustering with eigenvector selection , 2008, Pattern Recognit..

[5]  WangChih-Hsuan.  Recognition of semiconductor defect patterns using spatial filtering and spectral clustering , 2008 .

[6]  Santosh S. Vempala,et al.  On clusterings-good, bad and spectral , 2000, Proceedings 41st Annual Symposium on Foundations of Computer Science.

[7]  Zengyou He,et al.  Discovering cluster-based local outliers , 2003, Pattern Recognit. Lett..

[8]  Ulrike von Luxburg,et al.  A tutorial on spectral clustering , 2007, Stat. Comput..

[9]  Michael I. Jordan,et al.  On Spectral Clustering: Analysis and an algorithm , 2001, NIPS.

[10]  Chih-Hsuan Wang,et al.  Recognition of semiconductor defect patterns using spatial filtering and spectral clustering , 2008, Expert Syst. Appl..

[11]  Jitendra Malik,et al.  Normalized cuts and image segmentation , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.