On Kaggle you can find some competitions and download the associated datasets.
There is a system that scores your solutions in real time and you'll see your place on the "live leaderboard".
It's a good way of studying machine learning techniques because choosing a "for knowledge" competition you can compare your solution with other participants and discuss strengths and weaknesses of various approaches.
发布评论
评论（4）
在 Kaggle 上，您可以找到一些竞赛并下载相关数据集。
有一个系统可以实时对您的解决方案进行评分，您将在“实时排行榜”上看到自己的位置。
这是学习机器学习技术的好方法，因为选择“知识”竞赛，您可以将您的解决方案与其他参与者进行比较，并讨论各种方法的优点和缺点。
On Kaggle you can find some competitions and download the associated datasets.
There is a system that scores your solutions in real time and you'll see your place on the "live leaderboard".
It's a good way of studying machine learning techniques because choosing a "for knowledge" competition you can compare your solution with other participants and discuss strengths and weaknesses of various approaches.
试试我的博客 Vellum Information，其中有一些带注释的参考书目，整理了数据集和数据源：
http://velluminformation.com/2014/03/05/big-data-public-databases-an-annotated-bibliography/。
我有一份带有注释的各种可用数据源的参考书目。我在这里还提供了带注释的健康数据参考书目：
http://velluminformation.com/2012/05/19/free-online-public-data-sources-an-annotated-bibliography/。
明显的披露，这是我的博客，所以那里还有其他技术内容。
Try my blog, Vellum Information, where I've got several annotated bibliographies curating data sets and data sources:
http://velluminformation.com/2014/03/05/big-data-public-databases-an-annotated-bibliography/.
I've got an annotated bibliography of various data sources that are available. I've also got an annotated bibliography for health data here:
http://velluminformation.com/2012/05/19/free-online-public-data-sources-an-annotated-bibliography/.
Obvious disclosure, this is my blog, so there are other technical things on there as well.
UCI 机器学习档案 以及 KDD Cup 可能是最知名的通用数据挖掘档案。更具体的来源示例是 UCR 时间序列分类/聚类页面。
The UCI Machine Learning Archive and the past datasets of the KDD Cup are probably the best known such archives for general data mining. An example of a more specific kind of source is the UCR Time Series Classification/Clustering Page.
这是来自 DataWrangling.com 的一篇文章，其中列出了数百个数据集。
Here's an article from DataWrangling.com that lists hundreds of datasets.