previous | up | next       DiffMeasure class

DiffMeasure

Click on the image for a live example

Root mean square difference measure between the results of a calculation and data either from an earlier calculation or read from a file. There is a running display of the current difference measure continually updated as any parameter in the model is changed.

There are only two fields to be filled in, both of which are pointers which can reference objects anywhere in the current model. The first requires an LineSet object such as the results of a previous calculation saved with the save to clipboard option on the results display or read in from a file. All available instances of LineSet are shown on the DataClipboard. To make the connection, click the bottom left corner of an item on the data clipboard. Its name will then be dragged around by the mouse. Drag it over the data box on the DiffMeasureEditor and click again. As long as the object is compatible with what is required, its full name within the current object will appear in the box.

For the model-calc the desired object is dragged and dropped in just the same way, except this time it requires a CcmbCalc object and will refuse anything else. Find the object you want to use in the main CatacombEditor display. Click on it with the right button and select copy from the menu. This will pick up a reference to that object which can then be dropped in the model-calc field. Dropping it anywhere else will have no effect.

The easiest way to try out the DiffMeasure is to compare a calculation with results saved from the same model with slightly different parameters. Select save to clipboard from any results viewer; pick the results up off the clipboard and drop them on the data box. Then drop the calculation object that made this data on the model-calc object, and the display should begin registering the difference between the two. A new point will be added each time you change a parameter value. Note however, that the basic DiffMeasure is not very subtle. It only does a point by point difference, so if you change the timestep between the data and model it will be comparing points at different times.