Targeting business users with decision table classifiers

Business users and analysts commonly use spread-sheets and 2D plots to analyze and understand their data. On-line Analytical Processing (OLAP) provides these users with added flexibility in pivoting data around dierent attributes and drilling up and down the multi-dimensional cube of aggregations. Machine learning researchers, however, have concentrated on hy-pothesis spaces that are foreign to most users: hyper-planes (Perceptrons), neural networks, Bayesian net-works, decision trees, nearest neighbors, etc. In this paper we advocate the use of decision table classiers that are easy for line-of-business users to understand. We describe several variants of algorithms for learn-ing decision tables, compare their performance, and describe a visualization mechanism that we have im-plemented in MineSet. The performance of decision tables is comparable to other known algorithms, such as C4.5/C5.0, yet the resulting classiers use fewer at-tributes and are more comprehensible