Evaluating Hybrid LSTM-RF Models for Superior Stock Price Prediction in Volatile Markets

The stock market is an essential element of a nation's financial system, significantly influencing capital distribution and economic development. Precise stock price prediction models empower investors to make educated choices, enhance portfolio returns, and reduce risks linked to significant market volatility. The non-linear and dynamic characteristics of stock prices provide difficulties in developing dependable prediction models. This research examines the prediction of Tesla's stock closing prices by contrasting conventional statistical models with sophisticated machine-learning methodologies. We evaluate ARIMA, SVM, LR, RF, and LSTM (Linear Regression, Support Vector Machine, and Long Short-Term Memory) models and their effectiveness. We introduce LSTM-RF, a hybrid model that combines LSTM's sequential learning capabilities with Random Forest's ensemble robustness. According to the results of the evaluation, the LSTM-RF hybrid model achieves better performance on a number of metrics, such as MAE, RMSE, R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup>, and MARE. Results show that although standalone ML models outperform traditional statistical methods, the hybrid LSTM-RF model achieves better accuracy and durability, making it ideal for forecasting high-volatility stocks like Tesla.