The Core Cluster-Based Subspace Weighted Clustering Ensemble
Algorithm 2
Input: ,
Output:
1. Performed clustering on each base subspace to generate a set of subspace clustering ensemble
2. Generate the adaptive core clusters according to Definition 1
3. Calculates the average distance between core clusters according to Equations (8) and (9)
4. Calculates the CSI of the cluster according to Equation (10), and is achieved according to Equations (11)–(13)
5. Construct the core cluster similarity matrix () according to Equations (14) and (15), and the weighted core cluster similarity matrix according to Equation (16).
6. Build the graph with ,
7. Constructed the normalized graph Laplacian according to Equation (20).
8. Perform eigendecomposition on to achieve the first eigenvalues and corresponding eigenvectors to build
9. After normalizing , perform -means to categorized the core clusters
