Quantitative Forecasting of Moderna Stock Prices Using Linear Regression, Random Forest, and LSTM

Moderna, a trailblazer in mRNA-based biotechnology, has experienced substantial stock price volatility, particularly during the COVID-19 pandemic. Accurate prediction of such price movements is essential for effective decision-making and risk mitigation in financial markets. This study evaluates the predictive performance of three modeling approaches—Linear Regression, Random Forest, and Long Short-Term Memory (LSTM)—for forecasting Moderna’s next-day stock prices. The analysis incorporates key financial indicators, including short-and-long-term moving averages (MA10 and MA50), the Relative Strength Index (RSI), and the Moving Average Convergence Divergence (MACD), which capture both momentum-based signals and underlying trends. Among the tested models, LSTM demonstrated the highest predictive accuracy by effectively analyzing temporal dependencies and sequential patterns within stock price data. Random Forest exhibited strong adaptability to non-linear relationships, while Linear Regression offered an interpretable and computationally efficient baseline. The findings highlight the importance of adopting advanced machine learning models, such as LSTM, for analyzing volatile stock markets, particularly within industries characterized by rapid innovation and dynamic market conditions.

I. Introduction

Predicting stock prices has long been a cornerstone of financial research due to the complex and interrelated factors that influence market behavior. The ability to forecast price movements accurately is essential for optimizing investment strategies, mitigating risks, and navigating volatile market conditions. Traditional statistical models have often struggled to account for the non-linear dependencies and dynamic nature of financial time series. However, the advent of machine learning and deep learning techniques has introduced innovative solutions capable of identifying intricate patterns and dependencies within financial datasets [1], [2]. As a leader in mRNA-based biotechnology, Moderna Inc. exemplifies the intersection of advanced technology and financial volatility. The company’s AI-driven approach to drug development and vaccine production has positioned it at the forefront of medical innovation. During the COVID-19 pandemic, Moderna’s rapid vaccine rollout highlighted the transformative potential of its AI-backed strategies, contributing to significant fluctuations in its stock price. These dynamics make Moderna’s stock an ideal case for exploring the application of predictive modeling in volatile financial markets [3], [4]. Prior research has explored a variety of predictive techniques, each with distinct advantages and limitations. Linear Regression is widely regarded for its simplicity and interpretability, making it a reliable baseline. However, its reliance on linear assumptions often limits its applicability in capturing complex market behaviors [5]. Ensemble models like Random Forest offer greater flexibility in modeling non-linear relationships and feature interactions, but they can be prone to overfitting in highly volatile datasets [6], [7]. Deep learning approaches, particularly Long Short-Term Memory (LSTM) networks, excel at processing sequential data and leveraging temporal structures, making them highly effective for analyzing financial time series [8], [9].