Abstract
This study evaluates and compares the performance of ARIMA and XGBoost models in predicting the weekly closing prices of the stock of Saudi Telecom Company. Our findings suggest that XGBoost outperforms ARIMA across all evaluation metrics, highlighting the effectiveness of machine learning in forecasting stock prices. The study demonstrates the limitations of traditional statistical models in predicting stock price fluctuations and emphasizes the potential of machine learning techniques in uncovering hidden relationships and trends in data. Our research provides valuable insights into the suitability of different prediction models for stock price forecasting and highlights the need for continued exploration of machine learning techniques in finance.