|کد مقاله||کد نشریه||سال انتشار||مقاله انگلیسی||ترجمه فارسی||نسخه تمام متن|
|4977344||1367709||2018||10 صفحه PDF||سفارش دهید||دانلود کنید|
- The design of a fast iterative algorithm for blind decomposition of tensors into subtensors,
- with a inspiration of learning method particularly adapted for large data set and high dimensions.
- The decomposing of the tensor in a set of subtensors in the same way as factorial analysis. These subtensors constituting a kind of latent variable representation of the sources of the measured phenomenon (images or signals) which can be used further for pattern recognition, data compression, etc.
In this work we present a novel algorithm for nonnegative tensor factorization (NTF). Standard NTF algorithms are very restricted in the size of tensors that can be decomposed. Our algorithm overcomes this size restriction by interpreting the tensor as a set of sub-tensors and by proceeding the decomposition of sub-tensor by sub-tensor. This approach requires only one sub-tensor at once to be available in memory.
Journal: Signal Processing - Volume 144, March 2018, Pages 77-86