MATLAB implementation of A Nonlinear Orthogonal Non-Negative Matrix Factorization Approach to Subspace Clustering A non-negative matrix factorisation based unsupervised clustering algorithm applied to clustering of images (face identity recognition) and general numerical data.
Datasets: We have used the UCI Machine Learning Repository data sets and the Cambridge Computer Laboratory Database of Faces AT&T. The 5 UCI datasets are Soybean, Zoo, Glass, Dermatology and Vehicle. The AT&T face database consists of gray scale face images of 40 persons. Each person has 10 facial images under different light and illumination conditions and the images from the same person belong to the same cluster. The clustering accuracy is evaluated by the common clustering accuracy measure which computes the percentage of data points that are correctly clustered with respect to the external ground truth labels.
Acknowledgement: This work is supported by Croatian Science Foundation IP-2013-11-9623 "Machine Learning Algorithms for Insightful Analysis of Complex Data Structures", EU Horizon 2020 SoBigData project under grant agreement No. 654024, the European Regional Development Fund under the grant KK.01.1.1.01.0009 (DATACROSS) and the Croatian Science Foundation IP-2016-06-5235 "Structured Decompositions of Empirical Data for Computationally-Assisted Diagnoses of Disease".
Cite: When using the code in your work please cite "A Nonlinear Orthogonal Non-Negative Matrix Factorization Approach to Subspace Clustering" by Dijana Tolic, Nino Antulov-Fantulin and Ivica Kopriva, https://doi.org/10.1016/j.patcog.2018.04.029