The Relationships Among Various Nonnegative Matrix Factorization Methods for Clustering

Binary Matrix Factorization with Applications

“…We use accuracy as the clustering performance measure. Accuracy discovers the one-to-one relationship between clusters and classes and measures the extent to which each cluster contained data points from the corresponding class [16]. The experimental results are shown in Table 6.…”

“…The nonnegative matrix factorization (NMF) has been shown recently to be useful for clustering [16]. NMF can be traced back to 1970s (Notes from G. Golub) and is studied extensively by Paatero [22].…”

“…Table 5 summarizes the characteristics of the document datasets. More detailed description of these datasets can be found in [17,16]. To pre-process the datasets, we remove the stop words using a standard stop list, all HTML tags are skipped and all header fields except subject and organization of the posted articles are ignored.…”

A convergent algorithm for orthogonal nonnegative matrix factorization

“…But, because adjusting α for each different dataset is rather unfair, we believe that these are the best results can be offered by MU-U. All bi-orthogonal NMF algorithms, D-B, MU-B, and AU-B, performed rather poorly in these datasets, which was unfortunate since there are some works that show D-B is a better clustering method compared to LS and D-U [11,28].…”

Block-Iterative Algorithms for Non-negative Matrix Approximation

“…However, NMA became popular after Lee&Seung described simple alternating descent algorithms for it in [12]. Since then interest in this problem has literally exploded; we refer the reader to the articles [2,20,13] for extensive references.…”