This paper studies several interestingness and contingency table standardization to find a more robust algorithm instead of using support and confidence.