Comparative Analysis of ARIMA and BSTS Models for Electricity Price Forecasting in Major West Coast Cities

Abstract

This study presents a detailed comparative analysis of two prominent statistical models, the AutoRegressive Integrated Moving Average (ARIMA) and Bayesian Structural Time Series (BSTS), in the context of forecasting electricity prices in major West Coast cities of the United States, namely Los Angeles, San Francisco, and Seattle. Utilizing historical electricity price data obtained from the Bureau of Labor Statistics, the research meticulously trains and tests both models, aiming to evaluate their predictive accuracy and reliability in dynamic urban energy markets. The core objective of this research is to ascertain the effectiveness of ARIMA and BSTS models in the realm of electricity price forecasting, with a particular emphasis on their applicability in urban settings characterized by rapid market evolution. Through this analysis, the study offers valuable insights that could significantly influence energy market forecasting and policy formulation. The findings are particularly relevant for stakeholders in the energy sector, including policymakers and market strategists, as they navigate the complexities and challenges of modern energy landscapes. This research contributes to a more comprehensive understanding of statistical modeling techniques in energy market analysis and provides a foundation for more informed decision-making in energy policy and strategy development.