Research on the Classification Ability of Deep Belief Networks on Small and Medium Datasets
2013
Andrejs Bondarenko

Recent theoretical advances in the learning of deep artificial neural networks have made it possible to overcome a vanishing gradient problem. This limitation has been overcome using a pre-training step, where deep belief networks formed by the stacked Restricted Boltzmann Machines perform unsupervised learning. Once a pre-training step is done, network weights are fine-tuned using regular error back propagation while treating network as a feed-forward net. In the current paper we perform the comparison of described approach and commonly used classification approaches on some well-known classification data sets from the UCI repository as well as on one mid-sized proprietary data set.


Keywords
Artificial neural networks, deep belief networks, classification, restricted Boltzmann machines

Bondarenko, A. Research on the Classification Ability of Deep Belief Networks on Small and Medium Datasets. Information Technology and Management Science. Vol.16, 2013, pp.60-65. ISSN 2255-9086. e-ISSN 2255-9094.

Publication language
English (en)
The Scientific Library of the Riga Technical University.
E-mail: uzzinas@rtu.lv; Phone: +371 28399196