There's often complaints on posts like these that adding Machine Learning to any HN submission title will result in instant upvotes.
Yesterday, I built a deep learning model which can predict the optimal time to submit a post to Hacker News to maximize point score, and also predict the general probability of hitting the front page. In the context of this post, incidentally, the title is too verbose; both the point potential and and prob of hitting the front page is higher if only using the second half (w/o the explicit "Machine Learning" invocation): http://i.imgur.com/5vOTTXo.png
I'm kind of a python/ML newb, am curious to know: when doing ML stuff, do you guys usually do it interactively with Jupyter? The whole process of preprocessing, graphing, reducing dimensions, ML... do you prefer to do it one-by-one in Jupyter or Jupyter-like envrionment vs. just make one big python script? Just curious about common workflows.
I agree with this. If we're talking about the same thing, many of these are "Machine learning for hackers" type articles and only give a superficial exposure of machine learning with the implicit promise of quick and easy mastery. The problem is, there's really only one way to master this stuff and it's to open up a textbook and study hard. You really need to get a rigorous treatment of the theory to understand machine learning.
there are chapter endnotes and an appendix that looks very helpful.
meta-texts like this one are a must, because the subject is so broad. moreover, i find a lot of the oft-recommended books in this area very noisy on the theory side.
i imagine all of all of the available resources will improve in quality and become more standard over time.
I agree also I keep seeing so many articles, tutorials, guides, etc on how to do machine learning, neural nets, a.i., etc. It's like so... which one is the right one, or which one can actually get me from A to Z the easiest.
There's often complaints on posts like these that adding Machine Learning to any HN submission title will result in instant upvotes.
Yesterday, I built a deep learning model which can predict the optimal time to submit a post to Hacker News to maximize point score, and also predict the general probability of hitting the front page. In the context of this post, incidentally, the title is too verbose; both the point potential and and prob of hitting the front page is higher if only using the second half (w/o the explicit "Machine Learning" invocation): http://i.imgur.com/5vOTTXo.png