Entropy Regularized Unsupervised Clustering Based on Maximum Correntropy Criterion and Adaptive Neighbors

Constructing accurate similarity graph is an important process in graph-based clustering. However, traditional methods have three drawbacks, such as the inaccuracy of graph, vulnerability to noise and outliers, need for additional discretization process. In order eliminate these limitations, entropy regularized unsupervised clustering based on maximum correntropy criterion adaptive neighbors (ERMCC) proposed. 1) Combining information solve trivial distributions. And we introduce ℓ0-norm spectral embedding construct with sparsity strong segmentation ability. 2) Reducing negative impact non-Gaussian by reconstructing error using correntropy. 3) The prediction label vector directly obtained calculating sparse strongly connected components Z, which avoids Experiments are conducted six typical datasets results showed effectiveness method.