Feature Selection and Hyperparameters Optimization Employing a Hybrid Model Based on Genetic Algorithm and Artificial Neural Network: Forecasting Dividend Payout Ratio

Fatih Konak (), Mehmet Akif Bülbül () and Diler Türkoǧlu ()
Additional contact information
Fatih Konak: Hitit University
Mehmet Akif Bülbül: Nevşehir Hacı Bektaş Veli University
Diler Türkoǧlu: Hitit University

Computational Economics, 2024, vol. 63, issue 4, No 15, 1673-1693

Abstract: Abstract Among the most crucial factors that should be considered in the fundamental decision-making processes of companies is dividend policy. All market participants pay close attention to the decision of how much profit should be given or kept, as well as the assessment of the either the direct or indirect variables influencing the dividend that should be delivered, in order to maximize the shareholder’s present value. In light of the dividend payments made between 2011 and 2021 by the companies listed in the Borsa Ịstanbul Dividend Index, the research’s objective is to identify the internal or external factors that may have an impact on the dividend distribution rates of the companies and to establish a structure that can be used to project future dividend distribution rates. In this perspective, a feed-forward backpropagation artificial neural network model was used to forecast utilizing the 26 distinct input variables that were identified. Prior to this stage, a hybrid model was put out in a bid to enhance the efficiency of the Artificial Neural Network network structure’s parameters, which included input, hidden layer, number of neurons, and learning and activation functions in each layer. The hybrid structure employed allowed for the simultaneous optimization of the ANN’s network parameters and the feature selection from among 26 distinct parameters. The experimental tests decreased the number of characteristics that should be utilized to estimate the dividend distribution ratio to 5, and these 5 features were used to improve the optimum performance ANN network settings. Through experimental experiments using the hybrid model, successful results were attained with a mean square error of 0.075 in the most effective iteration. One may claim that the successful model proposed for the forecasting of dividend payments, which is anticipated to affect stock prices, is a collection of data that market participants can take into account.

Keywords: Dividend payout; Artificial neural networks; Genetic algorithm; IOS-Android application (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://link.springer.com/10.1007/s10614-023-10530-z Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:kap:compec:v:63:y:2024:i:4:d:10.1007_s10614-023-10530-z

Ordering information: This journal article can be ordered from
http://www.springer. ... ry/journal/10614/PS2

DOI: 10.1007/s10614-023-10530-z

Access Statistics for this article

Computational Economics is currently edited by Hans Amman

More articles in Computational Economics from Springer, Society for Computational Economics Contact information at EDIRC.
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().