Abstract. We propose a non-standard neural network called TPNN which offers the direct mapping from a peptide sequence to a property of interest in order to model the quantitative ...
Credit analysts generally assess the risk of credit applications based on their previous experience. They frequently employ quantitative methods to this end. Among the methods used...
Petri net faulty models are useful for reliability analysis and fault diagnosis of discrete event systems. Such models are difficult to work out as long as they must be computed ac...
Edouard Leclercq, Souleiman Ould el Medhi, Dimitri...
This paper proposes a novel constructive learning algorithm for a competitive neural network. The proposed algorithm is developed by taking ideas from the immune system and demons...
Helder Knidel, Leandro Nunes de Castro, Fernando J...
Results of neural network learning are always subject to some variability, due to the sensitivity to initial conditions, to convergence to local minima, and, sometimes more dramat...