Trend Price Prediction of the S&P500 Using the VIX index, Moving Average and Trading Volume in a Combined Random Forest and LSTM Model

Accurately predicting stock price movements, particularly for major indices like the S&P500, is a critical objective for investors and financial analysts to optimize portfolio performance and manage risk. This study investigates the effectiveness of machine learning models Random Forest (RF) and Long Short-Term Memory (LSTM) combined with technical indicators such as the VIX index, moving averages, and trading volume, to forecast price trends in the S&P500. A series of experiments were conducted using historical price data to evaluate whether these models, individually or in combination, could outperform traditional investment strategies such as buy-and-hold or random buy heuristics. The results show that the LSTM model, without incorporating the VIX index, achieved the highest F1-score of 68%, performing better than a random buy strategy. However, neither the LSTM nor the Random Forest models were able to outperform the buy-and-hold strategy. Additionally, integrating LSTM predictions as a feature into the Random Forest model did not enhance predictive accuracy. A notable finding is that shifting the testing periods negatively affected the models’ predictability. These insights highlight the limitations of current approaches, pointing to areas for improvement such as integrating the VIX index into LSTM models to enhance prediction accuracy.