Journal of Engineering and Applied Sciences
ISSN: Online 1818-7803ISSN: Print 1816-949x
Abstract
The portfolio management is a very important problem in the econometric field. In this research, we propose a new model by adding a new constraint to the Markowitz’s Models to avoid the investigation into the assets of a negative return. Because of its effectiveness, continuous hopfield network is used to solve the proposed models. In this regard, we construct an original energy function that makes a compromise between the risk, profit and cardinality constraints. To ensure the equilibrium point feasibility, the parameters of the energy function are chosen based on a consistence mathematical results; In addition, the slop of the activation functions is chosen such that the behavior of each neuron is almost leaner. We compare our method to several other ones, basing on real financial data. The proposed method produces the best solutions.
How to cite this article:
Kaoutar Senhaji, Karim El Moutaouakil and Mohamed Ettaouil. A Robust Neural Network Approach for the Portfolio Selection Problem
Basing on New Rational Models.
DOI: https://doi.org/10.36478/jeasci.2019.675.683
URL: https://www.makhillpublications.co/view-article/1816-949x/jeasci.2019.675.683