The Research Department of Economics and Management of Tadbir Nikan Business, Marketing, and Finance Open 3092-6238 In Press 2027 07 01 Designing a Model for Stock Price Prediction Based on the Integration of Machine Learning Models Designing a Model for Stock Price Prediction Based on the Integration of Machine Learning Models 1 21 EN Journal Article 2026 03 19 <p>This study aimed to design and evaluate an integrated machine learning model for predicting the stock prices of companies listed on the Tehran Stock Exchange using market, technical, trading, and financial predictors. This applied, quantitative, longitudinal, and predictive-modeling study was conducted on companies listed on the Tehran Stock Exchange during 2014–2024. After applying inclusion and exclusion criteria, 148 companies were selected as the final sample, and 301,284 firm-day observations were analyzed. The predictor variables included historical price data, trading volume, trading value, number of transactions, market capitalization, technical indicators, volatility indicators, and selected financial ratios. Data were cleaned, adjusted, normalized, and transformed into a supervised learning structure. The dataset was divided chronologically into training, validation, and test sets. Several machine learning algorithms, including multiple linear regression, ridge regression, decision tree, support vector regression, random forest, gradient boosting, extreme gradient boosting, and long short-term memory neural network, were trained and compared. Finally, an integrated stacking ensemble model was developed. The results showed that nonlinear machine learning models outperformed linear models in predicting adjusted closing stock prices. Among the individual models, extreme gradient boosting achieved the strongest performance, with MAE = 398.26, RMSE = 642.37, MAPE = 6.18%, R² = 0.948, and directional accuracy = 72.61%. The final stacking ensemble model demonstrated the best overall predictive performance, with MAE = 331.44, RMSE = 548.63, MAPE = 5.08%, R² = 0.964, and directional accuracy = 76.39%. Feature-importance analysis indicated that lagged adjusted price, moving averages, market capitalization, trading value, earnings per share, MACD, RSI, trading volume, volatility, and return on equity were the most influential predictors. The findings indicated that integrating multiple machine learning models improves stock price prediction accuracy compared with single-model approaches. The proposed stacking ensemble model can serve as an effective decision-support tool for investors, analysts, and portfolio managers in emerging stock markets.</p> https://www.bmfopen.com/index.php/bmfopen/article/download/492/358