Freely Available Datasets/APIs?

I've been studying algorithms about collective intelligence, for things such as recommendation engines, neural networks etc., and was looking for freely available datasets and/or APIs for practice projects. any suggestions?

EDIT: my original question was kind of vague, s开发者_Python百科o here are some examples of the types of things I was looking for:

Delicious API: collection of links, tags, who liked them, etc.

Ebay API: resource of various items, prices, descriptions.

Your question is pretty open-ended, so I don't know if this will help you, but...

Perhaps the largest single body of gathered AI knowledge was/is Cyc. You can download OpenCyc, which includes a data set of millions of hand-entered ontological assertions about the Real World (e.g. "water is wet").

There's also a metric boatload of code, both in C and Lisp, to work with that data.

Maybe the UCI Machine Learning Repository is what you're looking for.

http://archive.ics.uci.edu/ml/

Weka's website (a collection of Machine Learning algorithm) also features a list of datasets:

http://www.cs.waikato.ac.nz/ml/weka/index_datasets.html

Bit vague but I am about to release a neural network manager and framework written purely in PHP using MySQL and it comes with an API. If you want I could tell you where to download it, once I upload it.