An interestingness measure estimates the degree of interestingness of a discovered pattern and has been actively studied in the past two decades. Several pitfalls should be avoided in the study such as a use of many parameters and a lack of systematic evaluation in the presence of noise. Compression-based measures have advantages in this respect as they are typically parameter-free and robust to noise. In this paper, we present J-measure and a measure based on an extension of the Minimum Description Length Principle (MDLP) as compression-based measures for mining interesting rules. © 2009 Springer Berlin Heidelberg.
CITATION STYLE
Suzuki, E. (2009). Compression-based measures for mining interesting rules. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 5579 LNAI, pp. 741–746). https://doi.org/10.1007/978-3-642-02568-6_75
Mendeley helps you to discover research relevant for your work.