We present Nuclear Norm Clustering (NNC), an algorithm that can be used in different fields as a promising alternative to the k-means clustering method, and that is less sensitive to outliers. The NNC algorithm requires users to provide a data matrix M and a desired number of cluster K. We employed simulate annealing techniques to choose an optimal L that minimizes NN(L). To evaluate the advantages of our newly developed algorithm, we compared the performance of both 16 public datasets and 2 real psoriasis genome-wide association studies (GWAS), comparing our method with other classic methods. The results show that our NNC method consistently outperforms other methods due to its higher robustness and accuracy. In conclusion, NNC is an efficient method for clustering, which is especially better than k-means in most real datasets.
Downloads:
0 This Week