An attribute binning algorithm for association rule induction

Abstract:

Binning or discretization of continuous attributes is a preprocessing task required by many data mining algorithms. This task becomes especially important in association rule mining since it has a significant influence in the quality of the induced rules. In this paper, a multivariate method for binning quantitative attributes is proposed. Data from software projects was used for validating the method. Resulting discretized data used for mining the rules lead to a comprehensible set of interesting and high confident association rules, which cover a high percentage of records in the data set.

Año de publicación:

2007

Keywords:

• Association Rules
• Binning
• Discretization
• Clustering

Fuente:

scopus