Introducing the (S)ARIMA models

With our little recap of regression taken care of, we can start talking about the requirements and different components of the ARIMA and SARIMA models, including how they’re the same and how they’re different. However, before we get into the formula and the two variations on the regression that make up the (S)ARIMA model, let’s cover some of its requirements.

Requirements of the (S)ARIMA model

While the (S)ARIMA model is famously an effective option for forecasting time series data and requires far less data than many alternative approaches, it does come with a little baggage. Unlike the techniques we’d likely bucket into the Machine Learning category, such as neural networks or even regression forests, the (S)ARIMA model requires a few things for our underlying data distribution, namely that our data is stationary. Let’s dive into this topic a bit more.

What is stationarity?

The primary requirement of...