Fitting the model and generating forecasts
We’ve induced stationarity in our data, we’ve inspected the trend, we’ve investigated the seasonal patterns, and we’ve plotted the ACF and PACF. We’re ready to fit the model to our data.
The data
We’ll be working with hourly temperature data to deploy a SARIMA model and generate our forecasts. The following data is hourly temperature data from Los Angeles, California, taken in 2013:
Figure 7.13 – Hourly temperature data from Los Angeles, California
From here, we’ll move through the steps to explore our data and train an effective SARIMA model to generate a 5-day forecast of this temperature data. We will skip the data cleaning steps as they have been addressed in earlier chapters.
Does our data look stationary?
Recall the very first topic that we discussed under the (S)ARIMA section, stationarity. To deploy an (S)ARIMA model, we require weak stationarity...
