Incremental spectral clustering by efficiently updating the eigensystem
摘要: Various unusual conditions are likely to occur during sewage treatment process, which would lead to some consequences such as the decrease of water quality in the process of sewage treatment and the increase of disposal process, whereby causing a great influence to the practical operation efficiency of sewage treatment factories.
Finally, in order to solve the disadvantage of spectral clustering, some improvements are introduced briefly. By putting this algorithm into the fault classification in sewage treatment, the results demonstrate that this algorithm could be an effective identification towards the unusual conditions during the sewage treatment process, which provides an efficient way for sewage treatment process fault diagnosis. Based on the analysis of the fault characteristics during the process of active sludge sewage treatment, a PSO clustering algorithm is presented. Quantitative measures of change based on feature organization: Eigenvalues and eigenvectors. Proceedings CVPR'96, 1996 IEEE Computer Society Conference on. A min-max cut algorithm for graph partitioning and data clustering. ICDM 2001, Proceedings IEEE International Conference on.