O'Reilly logo

Practical Predictive Analytics by Ralph Winters

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Time series validation

Of course time series models also use this concept; however, time periods alone can be used to partition test and training data, for example, a model might be built on the previous five years of historical data (training) and then forecast on the current sixth year of data (testing or validation data).

Variations of this are common. A walk forward testing modeler will develop a time-series model on years 1-5, test the results on year 6, and then develop a similar model on years 2-6, and then test the results on year 7. This procedure would be repeated until there is no more data to test. In all these cases, the model is built on the in-sample data, and tested on out-of-sample data which is similar to k-fold validation. ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required