Output details
11 - Computer Science and Informatics
University of Huddersfield
An Information-Theoretic Foundation for the Measurement of Discrimination Information
<17> This paper is the first attempt to solve two long-standing open problems in data mining based on information theory: to explicitly interpret the amount of discrimination information conveyed by a term in practical applications, and successfully introduce the concept of the extent of semantic relatedness between two terms into scientific discussions. This paper establishes a unified and sound theoretical framework, clarifies potential problems arising repeatedly from many areas related to large-scale data analysis, and suggests feasible solutions. IEEE TKDE is a leading journal in Computing Science with an impact factor of 1.85 in 2010.