Research Article
Robust Graph Structure Learning for Multimedia Data Analysis
1: Input: the data matrix , two balance parameters and . | 2: Initialize: set and to be the identity matrices, random matrix , and sample similarity matrix , . | 3: Repeat | 4: Update the matrix | 5: Compute the matrix and the Laplacian matrix | 6: Update the matrix | 7: Update the matrix | 8: Update the matrix | 9: | 10: Until convergence | 11: Output: the graph matrix |
|