That's great. I'm glad to see most recent datasets have a time series classification. Should be good for time-series databases domain. At the end, it comes down to what questions you want answered with available data. Take this one for example.

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

I know that the Bay Area police used to analyze power load curves to figure out people growing weed :)

I'm getting started in ML and yea it's a great combinations of datasets to play around with! Sad for the weed growers tho :(

Ah yes, this is the classic go-to place for machine learning data sets. I implemented my first ML models using these files, especially the famous iris data set[1] which is a good one for beginners. It was so exciting to see my first multi-layer perceptron working (and to compare its results to the simple perceptron or decision tree, etc, on the same data).

[1]: http://archive.ics.uci.edu/ml/datasets/Iris

Staring out in ML here as well, and glad to know it's the goto place for starters/veterans :)

Proud to see my alma mater on HN. Zot zot!

Zot zot ftw!

Go eaters!

Zot zot ^^

UCI-ML The classical source for dataset selection papers.

For reference

http://web.archive.org/web/20061109231206/http://www.jmlg.or...

"Choose the data set." links to overview articles. Is there a link to actual datasets?

There is a satirical paper (look at names of the authors) on doing "Dataset selection".

The sarcasm is strong in this list :]

I love this so much. This is how I got my first intro to ML. I picked out the income data set andto my pleasant surprise, I got high predictive accuracy from a decision tree model in R. Gave me lots of confidence to try it out on other data sets.

A wonderful thing, I used it for a number of workshops in ML.