primaryobjects

An API like that is good for an introduction, but I think you'll get better results with a machine learning approach (my pet project http://www.sentimentview.com). When I was running tests with a baseline algorithm (just matching against positive and negative keywords), I saw results from 62% accuracy with the baseline to 80% with an SVM http://blog.sentimentview.com/post/59031004797/learning-curv....

This is a classic categorical problem for machine learning. I'm surprised so many suggestions have involved formulating some sort of clever algorithm like skin detection, colors, etc. You could certainly use one of those for a baseline, but I'd bet machine learning would out-score most human-derived algorithms.

Take a look at the scores for classifying dogs vs cats with 97% accuracy http://www.kaggle.com/c/dogs-vs-cats/leaderboard. You could use a technique of digitizing the image pixels and feeding to a learning algorithm, similar to http://www.primaryobjects.com/CMS/Article154.aspx.

I just gave it a try and submitted a program. I scored 64% accuracy. Currently in 4th place, but I'm sure that won't last for long. http://www.kaggle.com/c/dogs-vs-cats/leaderboard

Same here. I've also been heavily occupied with ML and AI for the past year or two. I think it's going to usher in the next phase of computers, we'll see. Computers are finally getting fast enough to do some interesting things with all of this data we're creating.

This is a timely article, as my current project deals with Twitter sentiment analysis http://www.sentimentview.com.

Although, I'm not classifying by specific emotion type as the article describes. In all, I'm seeing accuracy rates of around 80%. This is in comparison to a brute-force word list, which scored an accuracy of 62%. I'd be curious to see what kind of accuracy the researches in the article are achieving, with so many sub-division classifications. Not to mention, how did they derive their initial model?

I say this with humor, but this weekend, I took part in an indie 48-hour game-jam contest. I created the game, Fire Starter http://dummysoftware.com/games/firestarter

Created it with HTML5, targeting the desktop, rather than mobile. Didn't win, but had a lot of fun. That's kinda like the op's story compressed into a time-span of a few hours, sort of.

Btw, I recommend game-jams to anyone interested in indie game development. It forces you to create something quick, take decisive actions, and finish the game.

This sounds like genetic algorithm programming (see my other comment below https://news.ycombinator.com/item?id=6080492). Writing a program consists of defining the end state. The GA then runs through thousands of epochs, getting closer and closer to the end state (determined by a fitness score for each program), until the solution is found.

I was able to achieve automated programs for printing text, simple math operations, string manipulation, and conditionals. After that, programming the fitness methods started getting complicated.

I'm still hoping for computer programs to be written without humans http://www.primaryobjects.com/CMS/Article149.aspx

That was my initial experiment with self-programming AI, although ultimately, the fitness methods were starting to grow in complexity themselves.

SSL encrypts everything in the url after the ? so GET or POST matters little.

Couldn't you also serialize the resulting model to JSON (including weights/values), stick that JSON hard-coded into your source, and deserialize at runtime to execute the model? This is how I typically deploy neural networks and other machine learning models.

For example, http://colorbot.herokuapp.com stores the optimal model for the neural network in JSON, right in the code.

I was just going to comment on the similarity as well. I think these built-in HTML5 editors are great, especially when they include persistence. For the node.js devs, I made a wrapper for Create.js http://github.com/primaryobjects/contentblocks

WhaleEye SEO Spider - http://www.whalei.com

I created the same thing a year ago. I underestimated the CPU load and number of pages on many sites (especially e-commerce sites). It doesn't take many crawls to quickly fill a db.

I've been worried about this, ever since their acquisition. However, if they shutdown FeedBurner, there would be a LOT of broken links suddenly left on the web. Kind of like bit.ly or tinyurl suddenly disappearing. I imagine the public backlash from this makes it a bit more tricky for Google to drop FeedBurner, than it was to drop Reader.

I'm also taking the class. Just finished the logistic regression programming assignment. Great stuff. You can take those algorithms and kind of add a bit of magic to your software.

++[>++++++++++++++++++++++++++++++++++++<-]>.++++++.

It would be nice if there was a url parameter to auto-populate the source code and run it. Generated the above with http://goo.gl/aD67C

This looks similar to the genetic algorithm technique I used for the AI Programmer project http://www.primaryobjects.com/CMS/Article149.aspx

Genetic algorithms (and variants) can pick up some very interesting patterns for achieving the target fitness. I don't think this is necessarily anything new, and GAs have their limitations, but applying them to multiple NES games is impressive. I'm convinced that GAs will be an important part of future computing.

It's nice, but I'm more of a fan of direct inline editing on the web page via HTML5 contenteditable=true. The Aloha Editor http://aloha-editor.org does this very well. I've actually just recently added their editor to my node.js mini cms, called ContentBlocks http://contentblocks.herokuapp.com

Yep. https://github.com/primaryobjects/redant

Nice job. I made almost the same exact interface a few months back in node.js + mongodb http://red-ant.herokuapp.com

I've actually written a search crawler for clinicaltrials.gov. Their XML format is fairly convenient for parsing. I never thought for mining those types of data-points though. I suppose you could look at the study completion status and whether results are posted or not http://www.clinicaltrialweb.com

Instead of just posting the values to the server, why don't you pass the element IDs as well? You can then name your "link" elements with a key substring that matches the parent element. When you get your postback, you can match the "links" with their parent "step" via the key. This way, you don't rely on the index value in an array. For example:

<input id="step1" />

<input id="step1_link1" />

<input id="step2" />

<input id="step2_link1" />

<input id="step2_link2" />

Visual Studio 2012

Notepad++

Web Server (IIS, Apache, Node, etc)

Photoshop/Paint.NET

Chrome

Firefox

Internet Explorer

Safari

CDN links for jQuery http://jquery.com/download

CDN links for jQuery UI http://jqueryui.com (bottom of page)

jQuery Block UI http://www.malsup.com/jquery/block (modal popups)

Twitter Bootstrap http://twitter.github.com/bootstrap (base css)

Open Internet connection

Wow, 249 added. And a friendly mix of script injection, iframes, css, and image tags (had to disable the inline script tags because the redirects made people unable to see the data). Someone even changed the background image of the Webshell.IO interface itself.

It might be a good idea for the webshell guys to add some basic xss protection, but otherwise, I think we've proven this is a neat idea.

I played with it @ http://webshell.io/prototype/Njc5Z/7

It seems to work. This uses a REST service I recently wrote in node.js + mongodb.

Edit:

14 people click the link and hit Run, but no one changes the text? Come on. I wrote a second little app to track this at http://webshell.io/prototype/Njc5Z/8 just keep clicking Run to see it update, as people run the original app and hopefully enter their name or something.

Clickable link: http://sdappschallenge.com/submissions/6543-san-diego-school...

Unfortunately, many banking web sites require short passwords with no special characters. Likely, due to legacy databases and storage. I've even seen 6 character max restrictions. Or worse, the 4-digit pin ones (10,000 combinations yikes, I sure hope they have a limit and lockout for incorrect guesses).