Comparative Analysis of Forecasting Chevron's Crude Oil Stock Performance with Machine Learning Techniques

Abstract

The objective of this study is to predict the Chevron’s Corporation stock market performance by conducting a comparative analysis of contemporary and conventional machine learning approaches, with a particular focus on the CNN-LSTM and ARIMA models. Given the unpredictable characteristics of the crude oil industry, forecasting stock prices with precision has emerged as a pivotal dilemma for both investors and analysts. This research utilizes ARIMA, which is representative of conventional time series forecasting methods, and CNN-LSTM, which embodies the latest advancements in deep learning techniques, to address the intricacies associated with predicting stock prices in the energy sector. Through a comprehensive data preparation process and the application of sophisticated modeling techniques, this study aims to rigorously assess the predictive capabilities of both models in forecasting Chevron's stock prices. Traditional statistical analysis often relies on the ARIMA model as a benchmark, while the CNN-LSTM model seeks to identify the complex, non-linear patterns prevalent in financial market time series data. This research conducts a comparative evaluation of the two models, focusing on their accuracy, strengths, and limitations. The findings carry important implications for the realm of financial forecasting, shedding light on how modern deep learning techniques stack up against traditional approaches in predicting stock market movements. Beyond contributing to scholarly debates on financial prediction, this study also provides actionable insights for financial analysts.