Which of the following forecasting methods is specifically designed to capture both autoregressive and moving average properties, often applied to non-stationary time series data?

ARIMA is a powerful forecasting method used for time series analysis, especially when the data is non-stationary. The AR (AutoRegressive) component accounts for the influence of past values on the current value, while the MA (Moving Average) component captures the dependency between an observation and a residual error. The “I” (Integrated) component makes the series stationary by differencing. This combination makes ARIMA flexible and suitable for various types of time series data, such as financial and economic data, where both autoregressive and moving average processes are relevant. Option A (Exponential Smoothing) is incorrect as it focuses on smoothing time series data, not on combining autoregressive and moving average features. Option B (Moving Average) is incorrect because it averages out short-term fluctuations but lacks an autoregressive component. Option C (Seasonal Decomposition) is incorrect as it decomposes a series without modeling dependencies. Option E (Linear Regression) is incorrect because it assumes a linear relationship and doesn’t account for time-lagged dependencies.