Chapter 4 Evaluating modelling accuracy

Evaluating the performance of a model is essential in understanding a model’s strengths and weaknesses for particular data scenarios. Identifying the best model for the data is necessary in producing good forecasts.

In this chapter, we show some diagnostic tools which can be used to check that a forecasting method has adequately utilised the available information. Additionally, we discuss some accuracy measures that can be used to compare the forecasting performance across multiple models, and the difference in computing these measures on training data, with a test set, and using cross-validation.