Binary Information Press
There has been growing interest in uncertain data mining recent years, especially the uncertain data clustering. But few research works focus on the clustering analysis of uncertain data with arbitrary shape intrinsic features. For this problem, the authors propose an uncertain data clustering algorithm using multi-scale wavelet transform PCA, which called MSP-DBSCAN (Multi-ScaleWavelet Principle Component Analysis-Density Based Spatial Clustering of Application with Noise). In this algorithm, in order to discover intrinsic features with arbitrary shape clusters, firstly they pre-process uncertain data through wavelet transform and principal component analysis to extract features.