Journal of Management and Business Solutions 3092-7226 4 Serial Number 18 2026 03 10 Presenting a Memory-Instance-Based Gated Transformer (MIGT) Algorithmic Approach for Portfolio Management Presenting a Memory-Instance-Based Gated Transformer (MIGT) Algorithmic Approach for Portfolio Management 1 17 10.61838/jmbs.212 EN Journal Article 2025 10 27 <p>The objective of this study is to propose an efficient framework for portfolio management that can optimize investment returns while effectively controlling risk under both normal and highly volatile market conditions. The primary focus is on improving the accuracy of asset return forecasting. In this research, a memory-instance-based gated Transformer model is employed to predict asset returns. Financial data from the Iranian capital market covering the period 2016 to 2024 were collected and, after preprocessing, cleaning, and feature extraction, were entered into the modeling process. The mean absolute error in the test period was 0.0068, and the root mean squared error was 0.0100; in cross-validation, these values exhibited a standard deviation of less than 0.0004, indicating the stability of the forecasts. The paired Wilcoxon test used to compare the proposed model with benchmark methods produced statistics exceeding the significance threshold (p = 0.005), demonstrating a statistically significant improvement in the performance of the proposed model. The robustness of the results across different temporal subsamples and the preservation of acceptable performance under turbulent market conditions indicate that this approach possesses strong generalizability and practical applicability.</p> https://www.journalmbs.com/index.php/jmbs/article/download/212/171