Neural Networks New Capacity Factor Measurement for Improvement of SOM
Vahid Rahmati1, Morteza Husainy Yar2, Ali Reza Malekijavan3

1Vahid Rahmati, Currently Studying M.SC, Shahid Sattari University, Tehran Iran.
2Morteza Husainy Yar, Currently Studying M.SC, Shahid Sattari University, Tehran Iran.
2Dr. Ali Reza Malekijavan, Professor, Shahid Sattari University, Tehran Iran.
Manuscript received on 10 May 2014 | Revised Manuscript received on 20 May 2014 | Manuscript Published on 30 May 2014 | PP: 7-9 | Volume-3 Issue-12, May 2014 | Retrieval Number: L16340531214/14©BEIESP
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: The artificial neural networks have an important role in current life with higher expectations. The art of using these ANNs give us a good insight for problem solving. For example the applications in pattern recognition and regression are two areas in which ANNs are working well. Signal processing itself investigates broad ranges of ANNs. The purpose of this paper is, scanning connection routes of one and two layer networks that may be used as default structure in data replacing and signal analyzing. When a signal is considered as a variable by a problem solver, the problem solver chooses the best possible ANNs to solve it. But the way we ensure the high possibilities of reliability for these types of networks, while the compatibility is still needed is important. We present several factors to measure the capability of a specific network, for a specific problem the one like E-machine learning. Formal proofs for claim are provided as well. Finally we try to optimize Kohonen SOM using factor C.
Keywords: Artificial Neural Networks, Kohonen SOM, Machine Learning, ANN Connection Route.

Scope of the Article: Neural Information Processing