Feature-Assisted-Neighborhood-Smoothing

FANS.R

Input -- A: adjacency matrix; X: feature matrix; lambda: tuning parameter; screen (boolean): controls if feature screening is performed. Output: P_hat: estimated graphon values or probability matrix.

CV_for_FANS.R

Takes a list of lambda values (LAMBDA) as input. Outputs the CV error for each lambda.

Other related algorithms are in "Matlab Code" folder

Plots in the paper

visualize_graphon.R

Figure 1, 2 (top), 3, 10, 11

MSE_vs_lambda.R

Figure 2 (bottom), 4

repeatCompare.m (in "Matlab Code" folder)

Figure 5

realdata_results.R and "Matlab Code":

Figure 6, 7, 8, 9