US Semiconductor Stock Forecast Based on ARIMA and ETS Models

The US semiconductor industry is positioned as a key pillar in contemporary global manufacturing and technology. Predicting the semiconductor price index is critical in deriving investment decisions and evaluating policies. This study applies two time-series prediction models: ARIMA and Exponential Smoothing (ETS) in forecasting the US Semiconductor Industry price index using monthly data from 2015 to 2025. The author runs the time series data against the two models to evaluate which one offers a better prediction. The findings of the analysis demonstrate that the two models effectively capture the hypothesized growth in the semiconductor industry’s price index; however, the ETS model demonstrates superior forecasting accuracy with a lower root-mean-square error (RMSE = 10.68) relative to the ARIMA model (RMSE = 20.73). The findings of the study showed that exponential smoothing offers a more stable short-term forecasting framework, while ARIMA remains valuable for analyzing cyclical and seasonal fluctuations. Overall, this paper provides a quantitative perspective on the trajectory of the U.S. semiconductor industry within an evolving technological and macroeconomic landscape.