This study explores the application of machine learning in predicting stock market movements, focusing on the S&P 500 index. By utilizing the ensemble empirical mode decomposition method, we extract intrinsic mode functions from gold spot prices and the closing prices of the S&P 500 index to serve as input features for classification models. Through empirical comparisons, logistic regression emerges as the most accurate classifier, effectively predicting the direction of S&P 500 movements one day ahead. This approach aids investors in managing risk and optimizing asset allocation strategies. Additionally, we address challenges posed by nonlinear and nonstationary data, enhancing the predictive power of the models.