Cases are the examples or instances listed row by row in a data set. For users of this Excel application these may be projects, or locations or groups within projects
Cases can selected at two stages of analysis:
1. At the beginning: When the data set is imported.
Some cases may be deliberately left out with the intention that they will be used later as a “test” data set, to test the predictive power of the models developed with the portion of the data set currently being used. In the field of predictive analytics this is called Cross-Validation. The default percentage of test cases is usually 30%.
With EvalC3 cases can be left out either:
- Prior to cutting and pasting data into EvalC3 (Input Data) or
- When working within the Select Data view, by using the normal Excel filter function.
2. Towards the end: When predictive models have been identified that have satisfactory levels of performance.
See the Within-Case Analysis page for advice on case selection strategies that are appropriate at this stage.