Auto-SARIMA

This component uses a heuristic approach to analyze the target series and fit a (S)ARIMA model for forecasting with automatically configured hyper-parameters.

An interactive view is produced to help interpret and visualize the model used and forecasts generated.

Before using this component verify that there are no missing values in your target series. If there are, you can impute them externally with the missing value node.

Two types of predictions are computed:
1. Forecast: forecast of the given time series h periods ahead.
2. In-Sample Prediction: generates prediction in the range of the training data.

Note: This component requires a Python environment with StatsModels package installed. In this blog post we explain how to setup the KNIME

Python extension:
https://www.knime.com/blog/setting-up-the-knime-python-extension-revisited-for-python-30-and-20

Input ports

1. Type: Table
   Input Series
   Input table containing a column to be forecasted.

Output ports

1. Type: Table
   Forecast
   Future forecast values of target series and upper and lower bounds of 95% confidence interval.
2. Type: Table
   Insample Prediction
   Insample model prediction on input rows.
3. Type: Table
   Error Metrics
   Insample error metrics of output model.