Automatic Bezier-Based Model for One Predictor

This panel contains a model for one covariate/predictor based on two Bezier curves which can be automatically placed by HEMI 2.

Below is an image of a typical model (Coastal Redwoods againts Mean Temperature from the BioClim dataset). The lines are:

• Green - covariate histogram
• Red - histogram of the covariate values for each occurrence point
• Blue - the histogram of covariate values for each occurrence divided by the covariate histogram entry (this is the Fz/F or "raw" output from MaxEnt).
• Black - the current model.

All of the histograms have been scaled such that there maximum value is 1.0 for visiblity against the model.

When you select "fit" at the top of the dialog, HEMI 2 will move the control points (square boxes on the model curves) to the optimal placement based on the "Fit" settings. You can also move these control points yourself. As the points are moved, HEMI will compute statistics and display them below this panel.

You can also right click on the panel to access other options including:

• Copy as Raster to the Clipboard - copies a bitmap image of the model to the clipboard which can then be pasted into another application for publication.
• Export as Raster to File - export a bitmap image of the model to a file to import into another application for publication.
• Fit Curve to Data - this will automatically fit just this one model to it's covaraite data using the "Fit" settings
• Type - change the type of model from continuous to categorical. This will also change the settings that are available.
• Settings - additional settings for which curves are displayed and for fitting the model to the data.