Compute cell-to-cell similarity matrix by solving the following optimization problem via ADMM $$ min_{Z,E} ||Z||_* + lambda ||E||_{2,1}\\ s.t. X = XZ + E\\ Z'1 = 1\\ Z_{i,j} = 0 for (i,j) \in Omega$$
computeM(D, X, lambda)
| D | the unweighted KNN adjacency matrix |
|---|---|
| X | normalized sample vectors |
| lambda | the balance term between the rank of Z and the error |
a list containing the low rank approximation of X and manifold learning error