Folder structure in some repos seems like a game is being built.

This is awesome.

Not yet an organization here, but I've used scikit-learn for my internship project and thesis.

Something I implemented during that work (barely working) was the EM algorithm. I was wondering if EM is of any interest of the library.

Expectation Maximization is a very generic inference scheme that has many declination depending on the structure of the models it is applied to. For instance there we already have Gaussian Mixture Models that can be fitted using EM in sklearn. There is also an implementation of k-means which can also be interpreted as an EM algorithm.

Quick look feeling: too much text, too few images (tl;dr). Maybe adding a few more graphics could help (also lowering your bounce rate).

Congrats on the project! It's really interesting to see electrical engineering tools using web and modern concepts instead of the usual the nasty-windows-only-IDE packaging.

Word up! ;)

This is great to hear! I used a similar construction on my final project (electrical engineering) but left the appropriate mathematical proof (if any) into the Future Works section. The classification rule I found to be better was:

  c = argmax_{c \in C} (log(P(c)) + Sum_i (N_{d,d_i} \times log(P(d_i|c)))

where

  c - best class
  C - set of classes
  P(c) - prior prob. of class c
  P(d_i|c) - conditional of the word d_i on the class c and
  N_{d,d_i} - frequency of word d_i in document d

Excuse me for the possibly weird notation.

Virtually this expression means it places a frequency power upon the word probability, which is basically assuming (naively, I think) "words that are more likely to occur should be empowered".

This is awesome. Is anyone doing it already?

I found this after a quick search. http://givecamp.org/ . Not exactly the same idea, but it's on a good track.

I tried to use it for a while but gave up when I couldn't find a simple way of finding my twitter follows that also have a Subjot account. Did anyone tried/succeeded doing this?

This would be the url to use http://subjot.com/explore#friends

You can find it via Explore in the header and then People on the sidebar.

Thank you for your reply! ;)

It seems I don't have any twitter friends on Subjot, yet..

Congratulations on Subjot! I really like the idea. Are there plans for some kind of 'reputation' for people in specific subjects they talk about?

We don't have a solid design for that but we've definitely kicked around the idea. If I want to find the top jotters on NFL, it would be great to see some sort of ranked list.

I'm not sure but I think he meant creating an alias, such as git task. For instance, your commands could be:

git task list

git task new "Fix homepage styling"

That can be achieved if you put something like this on .gitconfig:

[alias]

task = !tigger

Oh awesome I didn't know about that I'll definitely look into it. Do you think people would be more likely to use it because they can type 'git' instead of 'tigger' (or whatever name it ends up being)?

It would definitely fit into git's naming convention for commands better. Whether or not that makes it stickier with users is probably an question only focus testing can answer :)

Is it me or it does not accept accents? I tried 'André' and server got sad.

It seems to die on that. I'll look into it and roll out a fix as soon as the traffic dies down a bit.

awesome =/

seems like python's urllib doesn't do unicode encoding properly.

fyi, this is fixed now.