Gradient Confidence Bounds?

[geoffrey hutchison]

I’ve seen a few plots, particularly for Gaussian Process Regression with a nice concept for confidence bounds using gradient color. Now, I don’t like the color scheme (e.g., I’d actually draw the mean line in black, and decrease the color intensity of the density)

But I think it’s really nice at demonstrating that the regression line is the best fit, and the confidence bounds have a range of confidence.

Would it be possible to implement something like this in DataGraph? In the meantime, I’ve been doing bootstrap fitting with Python and drawing ~250-500 translucent lines with Matplotlib. It kinda works, but I like DataGraph a lot more.

 

 

[jib]

This is quite nice, I’d also like to know if this can be done in DG

[tom lawton]

Hijacking slightly- I’d like to be able to plot data points as Gaussian blobs, as a more intuitive expression of their errors than box-and-whiskers…