Investment Management with Innovation in Neural Networks and Metaheuristic Algorithms

Mostafa   Sohouli  Vahed; Mohammad Ali Aghaei; Fariborz  Avazzadeh Fath; Ali Pirzad

doi:10.61838/kman.ijimob.3.5.3

Authors

Mostafa Sohouli Vahed Ph.D.Candidate Of Accounting ,Department Of Accounting ,Yasuj Branch,Islamic Azad University,Yasuj,Iran Author https://orcid.org/0000-0002-5006-7321
Mohammad Ali Aghaei Associate Professor,Department Of Accounting,Tarbiat Modares University,Tehran,Iran Corresponding Author https://orcid.org/0000-0001-9391-727X
Fariborz Avazzadeh Fath Assistant Professor,Department Of Accounting,Gachsaran Branch ,Islamic Azad University,Gachsaran,Iran. Author https://orcid.org/0000-0003-3952-5683
Ali Pirzad Assistant Professor,Department Of Management,Yasuj Branch,Islamic Azad University,Yasuj,Iran Author https://orcid.org/0000-0002-4412-9522

DOI:

https://doi.org/10.61838/kman.ijimob.3.5.3

Keywords:

Trading System Management, Neural Network Innovation, Metaheuristic Algorithms, Technical Analysis

Abstract

Objective: Considering the issue of selecting an optimal and desirable stock portfolio, which all investors, both individual and institutional, face.

Method: The purpose of the current research is to present trading systems with innovation based on neural networks and metaheuristic algorithms grounded in technical analysis. Therefore, the criteria affecting stock selection in technical analysis have been examined. Consequently, from among the companies listed on the Tehran Stock Exchange during the years 2011 to 2021, 135 companies were selected as samples through a systematic elimination method and analyzed using a combination of innovative neural network methods and metaheuristic algorithms.

Findings: The results have shown that such a trading system produces comparable or better results compared to Buy & Hold and other trading systems for a wide range of stocks even over relatively longer periods.

Conclusion: For future work, it is planned to focus on combining more technical parameters and using convolutional neural networks (CNN) or other deep neural network models.