Csiszár’s divergences for non-negative matrix factorization: family of new algorithms
ICA'06 Proceedings of the 6th international conference on Independent Component Analysis and Blind Signal Separation
| Hi-index | 0.00 |
We introduce a probabilistic extension of non-negative matrix factorization (NMF) by considering binary coded images as a probabilistic superposition of underlying continuous-valued elementary patterns. We provide an appropriate algorithm to solve the related optimization problem with non-negativity constraints which represents an extension of the well-known NMF-algorithm to binary-valued data sets. We demonstrate the performance of our method by applying it to the detection and characterization of hidden causes for failures during semi-conductor wafer processing. We decompose binary coded (pass/fail) wafer test data into underlying elementary failure patterns and study their influence on the performance of single wafers during testing.