zeratul

29 ms·

181.
by zeratul 15y ago
Europe bans SOPA and airport XRAY scanners and yet neonazi parties get seats in parliaments. Definitions of freedom and democracy in U.S. and Europe are incongruent. If you lived at least ten years on both continents you'd understand.
182.
by zeratul 15y ago
Congrats! You may wanna add viewport statistics to your GA code: http://wishfulcode.com/2011/11/02/measuring-viewport-size-wi...
183.
SOPA would never happen in Europe(arstechnica.com)
15 points by zeratul 15y ago|11 comments
184.
by zeratul 15y ago
One more idea. You could add a "system error button" or something similar that a user can click and tell you how your NLP/ML algorithms work. Then, use it for improving your models or maybe even implement active learning.
185.
by zeratul 15y ago
Good work. Here are couple suggestions: - The long list of articles at the bottom of the page seems excessive; why not just show 10 most recent? expand when needed; it just clogs the UI - Good HTML should read like any other good source cod
186.
by zeratul 15y ago
196 clones, 171 watchers, 23 forks; that's just in the first hour from release. Not bad, not bad at all. EDIT: Doom 1 source code seems dirty when compared to Doom 3.
187.
Big data for R (May 2010)(cybaea.net)
1 points by zeratul 15y ago|0 comments
188.
by zeratul 15y ago
I'm not sure. It seems that the route is: HN -> Twitter -> A swarm of bots sitting on Twitter When I re-posted my app on Sunday to get extra stats I saw the same thing. The Terminator Judgment Day happened but just on Twitter
189.
by zeratul 15y ago
Same goes for JS files. There is only on page that I'm serving. I also know that there is 10% of visits are returning "customers". So there should be around 90 css/js files served per 100 html serves but it's not. Few minutes after releasin
190.
by zeratul 15y ago
It's not a standard feature for the Google Analytics tool but I found code that adds viewport logging: http://beantin.se/post/354097397/track-viewport-google-analy... http://wishfulcode.com/2011/11/02/measuring-viewport-size-wi...
191.
by zeratul 15y ago
Thanks. There are many numbers in those charts that I don't understand. I was hoping that HN community will help: Why there are so many 405 (method not allowed) errors? I get this from, e.g., Twitterbot. Am I right? If someone GETs HTML and
192.
IE has 1% of market share ... on HN: "Show HN" aftermath
50 points by zeratul 15y ago|33 comments
193.
by zeratul 15y ago
I see the biggest value in the auto-generated Nutritional Information. Obesity is one of the major factors behind high healthcare costs in the U.S. If you can incentives image upload along the recipes, you could build a nice database with i
194.
by zeratul 15y ago
This might cause unintended consequences for the HN button: http://hnapiwrapper.herokuapp.com/ For example, I'm using one for my app: http://hnapiwrapper.herokuapp.com/button.html?width=120&... I wander if it will cause the button t
195.
by zeratul 15y ago
Do we know how many days need to pass before it's possible to re-post exactly same URL? It seems to me that HN has a ton of undocumented features.
196.
by zeratul 15y ago
Without looking under the hood I'd say there could be at least four reasons why they fail (based on what most of the NLP literature is lacking): - did not remove contradicting information from the training sets (two very similar vectors hav
197.
by zeratul 15y ago
I suspect that you are referring to what's called "hierarchical text classification". If so then any classifier can be used for that. And it's not a simple problem. I found it to be a good way to deal with unbalanced classes if you understa
198.
by zeratul 15y ago
Text mining: most of the time is spent on gathering the data, curating the data, and working with your annotators (domain experts). After that, you try a dozen or more ways to covert documents into a matrix format. Then, you try a dozen
199.
by zeratul 15y ago
This is invaluable comment. Changes my whole perspective on how to represent the data. If I could I would use all my karma here. Thank you EDIT: My mentor always says "graph or table has to stand on it's own" but that does not resonate with
200.
by zeratul 15y ago
I know, I know. It's amazing how many times I broke this rule. I do better in future. Somehow I think there should be a different design rules for labeling "persistent" (?) graphs, i.e. graphs that you look at regularly. If you look at a gr
201.
by zeratul 15y ago
Thank you very, very much. I will implement all your suggestions over the weekend. I was trying to make the app work with mobile devices, hence, skimpy use of the real estate. - If you click on the graph you get the definitions. - If you cl
202.
Show HN: When is good time to submit a story on HN?(hnpickup.appspot.com)
269 points by zeratul 15y ago|46 comments
203.
by zeratul 15y ago
Yes. That's what happened with the patent petition: a meaningless response. But then someone created another petition to re-consider the patent petition. Even if we have to go in circles at some point it has to stop. Media will blog about i
204.
by zeratul 15y ago
The URL link needs a token-hash at the end. I went through that when posting the link on HN some time ago: http://news.ycombinator.com/item?id=3206604 http://news.ycombinator.com/item?id=3198244 I didn't get much respond even though the
205.
by zeratul 15y ago
You have to be careful which countries can enforce U.S. based NDAs. You will probably need to translate it to a native language and have it notarized by a native lawyer and signed by the individuals (not by the company since that might not
206.
by zeratul 15y ago
I had to re-submitted this post. But in the original post there was a comment by amatus: "I started with telling Spock "xyzzy" and they quickly all started wishing each other "Merry Christmas". I only let the loop continue un
207.
Two Star Trek nerds talk with a woman - that never works, even for chatbots(sheepridge.pandorabots.com)
1 points by zeratul 15y ago|1 comments
208.
by zeratul 15y ago
Karpathy: Do you know if prof. Ng will cover the issue of sparse data in his lectures? When the number of dimensions is much greater than the number of samples and most of your cells in a matrix are equal to zero then most of the ML algorit
209.
by zeratul 15y ago
Looks promising: http://plexusengine.com/ I surely would like to hear more. Will that product become a data platform or app platform or both? The blog says: "an app and data platform". Will I be able to plug my own algorithms there?
210.
by zeratul 15y ago
Wallawe: Do you think this on-line petition system is useful? Instead of going through a congressman, one could go through the White House? https://wwws.whitehouse.gov/petitions#!/how-why/terms-partic...

More ›