can't reproduce the experimental results

[cqfei]

I can't reproduce the experimental results in the paper “Robust Graph Learning From Noisy Data” with this code, and the dataset in question includes JAFFE, COIL20, BA, TR11, TR41, TR45. I don't know what happen?

[cqfei]

I run the file Clustering_main.m and set the parameters alphalist=1/sqrt(max([m n])), betalist = [1 5 10 15 20], mulist = [1 5 10 15 20] to find the best result, which is far away from the result shown in TABLE II in the paper

[cqfei]

I suspected the problem was in updateS.m. There was a comment out in the file. Is this code useful? I've uncommented this code and it still doesn't work. What's wrong? Is this code reliable?