Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Avis-Mergulus]

Today in (high school senior) programming class, we went over so-called 'netiquette'
Teacher said the one of the important things to not do was post in supersize letters, all caps, etc

Why do schools insist on making a mockery out of informatics/IT? It shouldn't be about fuckin' "netiquette" or whatever senile concept some clueless fart somewhere considered necessary. It should be the beginnings of system architecture, an introduction to algorithms, and some programming assignments in a halfway sensible language, not... "netiquette". Whatever the hell that even is.

What is it?

It's based on a relatively less well-known machine learning method that I've found recently, something like a decision tree, but with intersections between branches. They use a very complicated method to construct their "decision jungle", but I've thought of a much simpler version that would, hopefully, retain the beneficial properties of such system. Hopefully.

That sounds publishable af, honestly. Try a basic implementation; sample datasets for testing can be found here, or here, or on Kaggle. If that's what you meant by "nothing to test it on".

[IcyTea31]

Yeah, netiquette is a bit outdated and not really worth spending too much time on. It can be summarized in a single sentence: "Don't be a dick." If that's a no-brainer to you, then it needs no further attention. If it doesn't feel like a no-brainer, you're going to need something more than a list of rules, and that something oughtn't be taught in a class on IT.

[Sergarr]

What is it?

It's based on a relatively less well-known machine learning method that I've found recently, something like a decision tree, but with intersections between branches. They use a very complicated method to construct their "decision jungle", but I've thought of a much simpler version that would, hopefully, retain the beneficial properties of such system. Hopefully.

That sounds publishable af, honestly. Try a basic implementation; sample datasets for testing can be found here, or here, or on Kaggle. If that's what you meant by "nothing to test it on".

Well, it sounds cool, but so was the last 20 or so times when I thought "this looks cool" and it turned out that I've either invented a bicycle or it just plain didn't work out in practical implementation upon further consideration. So I'll remain sceptical.

Also yeah that's what I've meant. I know that there is a lot of stuff like that in R, too. The problem is, writing the algorithm out in R would be a task full of pain. Local implementations of decision trees are weird looking monsters that rely heavily on external .cpp code, so I can't really use them as a base. Hell, I only wrote it in Warcraft 3, and only for a very, very specific reinforcement learning purpose.

Also x 2 combob, I really doubt that anyone would publish an article on machine learning from a "material science and technology of materials" person, kek.

[Avis-Mergulus]

What is it?

It's based on a relatively less well-known machine learning method that I've found recently, something like a decision tree, but with intersections between branches. They use a very complicated method to construct their "decision jungle", but I've thought of a much simpler version that would, hopefully, retain the beneficial properties of such system. Hopefully.

That sounds publishable af, honestly. Try a basic implementation; sample datasets for testing can be found here, or here, or on Kaggle. If that's what you meant by "nothing to test it on".

Well, it sounds cool, but so was the last 20 or so times when I thought "this looks cool" and it turned out that I've either invented a bicycle or it just plain didn't work out in practical implementation upon further consideration. So I'll remain sceptical.

Also yeah that's what I've meant. I know that there is a lot of stuff like that in R, too. The problem is, writing the algorithm out in R would be a task full of pain. Local implementations of decision trees are weird looking monsters that rely heavily on external .cpp code, so I can't really use them as a base. Hell, I only wrote it in Warcraft 3, and only for a very, very specific reinforcement learning purpose.

Also x 2 combob, I really doubt that anyone would publish an article on machine learning from a "material science and technology of materials" person, kek.

Well, you can always submit your thesis for whatever conference is next, and see where it goes from there? Not like there are formal background requirements for that stuff.

I will continue recommending Python for its dataset/table library (Pandas), and the usefulness of the scikit-learn data preprocessing tools, even if you are making your own classifier from scratch. From there the task is sort of the same as in R, just less of a pain in the ass, cause, well, R. It does have a lot of datasets included with it, but that's probably not a good enough reason to choose it over other things.

[Sergarr]

R code doesn't require compiling, though, which makes it significantly faster for testing things out. And god knows, anything I write require looooots of testing before it starts to even just imitate working correctly.

[Avis-Mergulus]

R code doesn't require compiling, though, which makes it significantly faster for testing things out. And god knows, anything I write require looooots of testing before it starts to even just imitate working correctly.

Eh, neither does Python... and my experience with compiled languages is limited to C#, which might not be a typical case because managed code and all that jazz, but I've never had a problem with debugging/testing compared to Python, and the build time is usually negligible unless you're building something dreadfully huge, which you probably aren't going to be, right?

I think you should just Shia LeBeouf it, with whatever tools. If you have the time/wish to do it, that is? You never know, it might turn out to be one of those "simplicity of genius" things. But no pressure? Hell, I dunno if I have the ability to exert pressure of any sort on you, but no pressuring is intended either way.

[IcyTea31]

I took a selfie for the first time in my life. I've started growing a beard, and want to chronicle my adventure for future generations.

[TheBiggerFish]

R code doesn't require compiling, though, which makes it significantly faster for testing things out. And god knows, anything I write require looooots of testing before it starts to even just imitate working correctly.

Well how long are you thinking it would take to compile?

But seriously go for it.

[itisnotlogical]

New Pandyland today. It was sadly a Skip Tobey comic, and not a very funny one at that, but all my webcomics have updated recently (relative to their own schedules) and been confirmed alive. _{except Prequel haha}

[IcyTea31]

Yolk was still runny

I honestly don't understand why people cook their eggs like that. In my experience, it only leads to half the yolk dripping onto the plate and the rest having a bland taste and terrible texture.

[TD1]

It's better with toast, though.

[TempAcc]

Runny yolk is the best yolk.

[scriver]

I took a selfie for the first time in my life. I've started growing a beard, and want to chronicle my adventure for future generations.

Post it!

[Avis-Mergulus]

I took a selfie for the first time in my life. I've started growing a beard, and want to chronicle my adventure for future generations.

Post it!

Yeah, I mean, we have a photo share thread for a reason, right?

[TD1]

Twist it! Bop it!