KDnuggets : News : 2002 : n20 : item5    (previous | next)

Features

From: Ronny Kohavi
Date: 13 Oct 2002
Subject: Three Real-world datasets available for association benchmarks


We are pleased to announce the availability of three real-world association datasets, one of which was used in the KDD Cup.
To use the data, you need to go through a simple click-through agreement.
The POS dataset provides over 3 million purchased items in over 500,000 transactions.

Note that these datasets provide item numbers only, not their description, hence their value is mostly for benchmarks (e.g., they are not useful for pruning rules to improve comprehensibility).

The datasets are available at http://www.ecn.purdue.edu/KDDCUP/index.html#assoc

For more details about the datasets, see
   Zijian Zheng, Ron Kohavi, and Llew Mason, Real World Performance of Association Rule Algorithms, KDD 2001.
available at http://robotics.Stanford.EDU/users/ronnyk/realWorldAssoc.pdf

  -- Ronny


KDnuggets : News : 2002 : n20 : item5    (previous | next)

Copyright © 2002 KDnuggets.   Subscribe to KDnuggets News!