International Journal of Computational Intelligence Systems Vol. 14(1), 2021, pp. 1763–1772 DOI: https://doi.org/10.2991/ijcis.d.210601.001; ISSN: 1875-6891; eISSN: 1875-6883 https://www.atlantis-press.com/journals/ijcis/ Research Article An Intelligent Hybrid System for Forecasting Stock and Forex Trading Signals using Optimized Recurrent FLANN and Case-Based Reasoning D. K. Bebarta 1 , T. K. Das 2 , Chiranji Lal Chowdhary 2,* , Xiao-Zhi Gao 3 1 Department of Information Technology, GVPCEW, Vishakhapatnam, India 2 School of Information Technology and Engineering, VIT, Vellore, India 3 School of Computing, University of Eastern Finland, Kuopio, Finland ARTICLE INFO Article History Received 03 Oct 2020 Accepted 06 May 2021 Keywords Stock forecasting, Dynamic time window, Recurrent FLANN, Firefly algorithm ABSTRACT An accurate prediction of future stock market trends is a bit challenging as it requires a profound understanding of stock technical indicators, including market-dominant factors and inherent process mechanism. However, the significance of better trading decisions for a successful trader inspires researchers to conceptualize superior model employing the novel set of techniques. In light of this, an intelligent stock trading system utilizing dynamic time windows with case-based reasoning (CBR), and recurrent function link artificial neural network (FLANN) optimized with Firefly algorithm is designed. The idea of using CBR module is to offer a dynamic window search to assist the recurrent FLANN architecture for superior fine-tuning the trading operations. This integrated stock trading system is intended to pick the buy/sell window of target stock to maximize the profit. To demonstrate the applicability of the projected system, the time-series stock data from IBM, Oracle and in currency Euro to INR and USD to INR exchange data on daily closing stock prices are used for simulation. The performance of the proposed model is assessed using error measures such as mean absolute error and mean absolute percent error. Furthermore, the experimental results obtained with/without using CBR is exhibited for different stock and Forex trading data. © 2021 The Authors. Published by Atlantis Press B.V. This is an open access article distributed under the CC BY-NC 4.0 license (http://creativecommons.org/licenses/by-nc/4.0/). 1. INTRODUCTION A significant portion of gross domestic product (GDP) is sourced from the corporate sector in the developed countries since the USA accounts for 70% of GDP from earnings of corporate revenue. Even the corporate in developing countries like India contribute a chunk of 15% toward its GDP. After globalization, majority of companies from the domicile have started expanding their businesses globally and referred to as multinational companies (MNCs). These com- panies raise funds by offering their shares to the general public through the open market. Consequently, financial market has been expanded and becomes a lucrative avenue for investment. People can buy the share, hold them, or sell them as per their convenience, and the difference in buying and selling price is the profit earned. To maximize the earning, every customer wait for the right time to buy/sell a particular portfolio. The trading opportunities are aggra- vating with an exponential increase in the number of companies listed under the ambit of stock exchange. Further, the advancement of information technology, especially in the finance and securities domain, provided the growth impetus of * Corresponding author. Email: chiranji.lal@vit.ac.in the financial sector leap and bound. In the last two decades, most of the trading is done electronically. Trading can be executed from an electronic gadget with an Internet connection due to which a stockbroker has rendered on-line trading services. Currently, most of the subscribers buy-and-sell stocks online [1]. With the advent of stock markets, people have the option to have multiple avenues to make their asset grow by investing in several funds like mutual funds, hedge funds, and index funds accord- ing to their financial profile. However, returns from these funds bank on the stock market performance only. Additionally, the stock market is also preferred by the Governments for venturing a part of their provident funds, retirement funds, and citizens’ saving to achieve attractive returns for their inhabitants. As a result, the finan- cial markets have evolved rapidly expanding into an interconnected global marketplace. The stock market is extremely volatile and continuously fluctuat- ing. As a result, the stock price of a company is unsteady. Stock indices forecasting is a complex and challenging task as it involves analyzing many factors. Despite this, the inherent capability of artificial intelligence, soft computing, machine learning (ML), and data mining [2], the complicated task of future stock price predic- tion becomes somehow practicable. However, in the face of several