search

JuliaML / MLPlots.jl

Plotting recipes for statistics and machine learning using Plots.jl
Other
24 stars 5 forks source link

Error with the ROC curve #8

Closed diegozea closed 8 years ago

diegozea commented 8 years ago

@tbreloff I'm using an updated version of matplotlib and the dev branch of Plots.

image

tbreloff commented 8 years ago

I haven't ignored this... I'm in the middle of making a pass through MLPlots to get everything working... but I got sidetracked because I also needed some fixes in OnlineAI, which was part of the MLPlots tests. I expect to make a big push with MLPlots today, and probably try to register the package in METADATA.

tbreloff commented 8 years ago

Question... should this be using the repo https://github.com/davidavdav/ROCAnalysis.jl or https://github.com/diegozea/ROC.jl?

Do you have a good (simple) example to be used for both MLPlots testing and an example on the readme?

diegozea commented 8 years ago

Hi @tbreloff ! This should use ROCAnalysis instead of my unregistered ROC. However the ROCR example in my package is good for testing this ;)

diegozea commented 8 years ago

@tbreloff The shadow under the ROC curve should be from x instead of from y=x.

diegozea commented 8 years ago

x means y=0

tbreloff commented 8 years ago

So just to be clear, you want fillrange=0? Check: https://github.com/JuliaML/MLPlots.jl#rocanalysis

diegozea commented 8 years ago

Exactly, It's fine now in the example :)