A min-max cut algorithm for graph partitioning and data clustering. ICDM 2001, Proceedings IEEE International Conference on.
Finally, in order to solve the disadvantage of spectral clustering, some improvements are introduced briefly.
Quantitative measures of change based on feature organization: Eigenvalues and eigenvectors. Proceedings CVPR'96, 1996 IEEE Computer Society Conference on.
Reasonable learning sequence helps to strengthen the knowledge reserve of the classifier.
Abstract: The learning sequence is an important factor of affecting the study effect about incremental Bayesian classifier.
Unlike the traditional stream data, these applications require incremental algorithms to handle not only insertion/deletion of data points but also similarity changes between existing points.