randomWalkByMatrixInv,matrix-method {netSmooth} | R Documentation |
The closed-form solution is given by f_ss = (1 - alpha) * (I - alpha * A)^-1 * f_0 and is computed by matrix inversion in this function.
## S4 method for signature 'matrix' randomWalkByMatrixInv( f0, adjMatrix, alpha, normalizeAdjMatrix = c("rows", "columns") )
f0 |
initial data matrix [NxM] |
adjMatrix |
adjacency matrix of graph to network smooth on will be column-normalized. |
alpha |
smoothing coefficient (1 - restart probability of random walk) |
network-smoothed gene expression