Here is some good links for some
public datasets for testing data mining algorithms:
All the datasets of the previous
SIGKDD KDD-CUP. Some of them are very large!
The UCI KDD Archive that contains around 40 larges datasets of different types.
The
UCI Machine Learning repository offers 185 public datasets.
The
Frequent Itemsets Mining Datasets Repository provides many classical datasets for frequent itemset mining. It also includes several implementations of classical algorithm such as Apriori and others for mining frequent itemsets, frequent closed itemsets, maximal itemsets, etc.
A few datasets and implementations of data mining algorithms can also be found on the
web page of M. Zaki.
A lots of links to free datasets :
http://www.kdnuggets.com/datasets/index.html