Quality Measures in Pattern Mining

Authors: Sebastián Ventura José María Luna

Publish Date: 2016

Volume: , Issue: , Pages: 27-44

PDF Link

Abstract

In this chapter different quality measures to evaluate the interest of the patterns discovered in the mining process are described Patterns represent major features of data so their interestingness should be accordingly quantified by considering metrics that determine how representative a specific pattern is for the dataset Nevertheless a pattern can also be of interest for a user despite the fact that this pattern does not describe useful and intrinsic properties of data Thus any quality measure can be divided into two main groups objective and subjective quality measures Whereas objective measures describe statistical properties of data subjective quality measures take into account both the data properties and external knowledge provided by the expert in the application domain

Keywords: