website content about juliaml.github.io HOT 9 CLOSED

I started us off with a base website, using JuliaStats as a rough template. Content needed!

from juliaml.github.io.

Can we agree on the packages we'd like to highlight here? I'm going to start a list in the first comment... please post anything that is missing. Note: we don't need to actually move these packages into the org... I just want to have them on the website (with links to the github pages) so it's all consolidated.

from juliaml.github.io.

Now would also be a good time for me to start work on MLPlots.jl (probably instead of StatPlots.jl, since no one showed any interest in that, and then I can combine things like ROC curves, etc)

I think such a package would be very important at some point. I am a little reserved at the moment because I don't want to dillude my attention from KSVM too much (I am already kinda on a tangent with the loss functions in LearnBase) after that I would most certainly contribute the SVM plots to it.

ps: It's not set in stone yet, but it looks like I will give an hour long "Introduction to ML" talk at the end of this month in which I will use Julia examples to explain basic concepts such as linear/logistic regression, penalties, loss functions, gradient descent and such. It might be the case that I'll have to implement some common plots for that, which I would gladly contribute

from juliaml.github.io.

Cool... feel free to use anything from ExamplePlots... I presented at the last julia meetup in NYC and I have lots of stuff here... some of it might be useful.

from juliaml.github.io.

https://github.com/antinucleon/XGBoost.jl

from juliaml.github.io.

@tqchen
Now it is part of DMLC: https://github.com/dmlc/XGBoost.jl

from juliaml.github.io.

@tbreloff could you go ahead and establish MLPlots.jl? I would kick it off by creating plots for the loss functions. I think it should live at your repo until it's mature enough to move to JuliaML.

from juliaml.github.io.

Working on this now. Initial commit is up... I'm working to make sure the corrplot works with the example in the readme.

https://github.com/tbreloff/MLPlots.jl

from juliaml.github.io.

Ok it's ready for contributions!

from juliaml.github.io.