Icon

03_​SARIMA_​Models

SARIMA Models - Exercise (Solution)

This workflow predicts the residual of time series (energy consumption) by seasonal autoregressive integrated moving average (ARIMA) models that aim at modeling the correlation between lagged values and controling for seasonality in time series.

URL: Slides on the KNIME Website https://www.knime.com/form/material-download-registration

Time Series Analysis
03. SARIMA Models

Summary:
In this exercise we'll train and score our first SARIMA models.

Instructions:
1) Run the workflow up through the Decompose Signal component, we’ll start this exercise from here

2) Partition the data using the Partioning node. Let’s use an 80/20 split. Make sure you check the box to take data from the top. This is important with time series data.

3) Apply the SARIMA Learner to the output from the Differencer node. Note that that the SARIMA node can take some time to run so keep the settings low for now.

4) Use an SARIMA Predictor node after the learner, you can configure the number of values you want to forecast here.

5) Attach the Forecast output from the SARIMA Predictor to the top port of the scoring metanode and the other half of our Partitioning node to the bottom. Run the scoring metanode and look at the results. Try this with different numbers of forecasted values. Do the scores change?

6) Analyze the residuals of the SARIMA model with the Residual Analyzer node. What can you say about the residuals?


Data Loading
Data Preparation
Model Training (Signal)
SARIMA Learner
Missing Value
Energy usage data
CSV Reader
RowID
Line Plot (Plotly)
Residual Analyzer
SARIMA Predictor
RowID
Numeric Scorer
Column Filter
Partitioning
Date&Time Aligner
convert date/time into Date&Time objects
String to Date&Time
Joiner

Nodes

Extensions

Links