Designing a Model to for Selection of Effective Variables of Forecasting Future Dividend of the Firms Member Tehran Stock Exchange

Makvandi, Payam; Jasbi, Javad; Alavi, Hasam

Designing a Model to for Selection of Effective Variables of Forecasting Future Dividend of the Firms Member Tehran Stock Exchange

Authors

¹ Faculty Member of I. A. U. Karaj Branch

² Associated Professor of I. A. U. Science and Reseach Branch Department

³ Faculty member of I. A. U. South Tehran Branch

Abstract

Decision making is on the most important roles of a manager. Meanwhile, selection of
effective input variables on decision making or forecasting problems, is one of the most
important dilemmas in forecasting and decision making field. Due to research and
problem constraints, we can not use all of known variables for forecasting or decision
making in real world applications. Thus, in decision making problems or system
simulations, we are trying to select important and effective variables as good data.
In this paper we proposed a hybrid model of Genetic Algorithm (GA) and Artificial Neural
Network (ANN) to determine and select effective variables on forecasting and decision
making process. In this model, genetic algorithm has been used to code the combination of
effective variables and neural network as a fitness function of genetic algorithm. The
introduced model is applied in a case study to determine effective variables on forecasting
future dividend of the firms that are members of Tehran stock exchange.

Keywords