Abstract
The study aimed to forecast stock prices and price movements in the Damascus Securities Exchange using advanced artificial intelligence models, specifically Long Short-Term Memory (LSTM) networks. It employed two groups of technical variables to achieve this objective: the first group included closing and opening prices, along with short- and long-term Exponential Moving Averages (EMA), to predict the next day’s stock price, while the second group used the Moving Average Convergence Divergence (MACD) indicator to forecast the future direction of price movements. The empirical application was conducted on the DWX index as a representative of the entire study population, based on daily data covering the period from 2019 to 2022.
The study reached significant findings that are consistent with global trends in financial forecasting using artificial intelligence. The implemented LSTM model exhibited very high predictive accuracy, as the arithmetic means of the predicted values (DWX closing values and MACD values for the test period) were almost identical to the arithmetic means of the actual realized values, reflecting the model’s ability to capture complex patterns in financial time series. These results are in line with recent international evidence, including empirical studies on the S&P 500 and emerging markets, which report prediction accuracies exceeding 90% for LSTM-based models when combined with technical indicators such as EMA and MACD
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.