A Comparative Analysis of the Latest Trends in Stock Market Forecasting

Abstract

Forecasting Stock market prices is ongoing research with researchers aspiring to develop models that supersede previous accuracy rates. Historically, statistical models like the GARCH and ARIMA models have been used for predictive analytics tasks, however, the volatility and complex nature of stock market data make it difficult to complete prediction tasks with such models. Various machine learning models like the Random Forest algorithm and Support Vector Machine have also been used for stock market prediction, at times achieving significant prediction accuracy in some tasks. However, this may not be sufficient. Of late researchers are experimenting with deep learning algorithms for stock market prediction with these remarkably improving accuracy rates. In this study, we conduct a comparative analysis of prominent stock price forecasting models to identify the best-performing model. By using a systematic review of the literature, we evaluate the top-performing models in terms of prediction accuracy, adaptability, and robustness. Results show that variations of the ensemble LSTM model outperform other models. Additionally, we observe that models with knowledge retention generally outperform traditional machine learning approaches in stock market prediction tasks. Our findings have significant implications for developing effective stock market prediction systems and contributing to the growth of AI-driven financial analytics.