A Survey of Semantic Relatedness Measures

Abstract

Semantic relatedness measures are widely employed today to estimate the strength the semantic relationship among two terms of two pieces of texts. In other words, the ability to compute semantic relatedness among terms has widespread applications and is considered as an important pre-processing step in Natural Language Processing. Despite an extensive amount of work dedicated to this area of research, the strengths and weaknesses of various methods have not been compared extensively. In this paper, a comprehensive survey of the broad notion of semantic relatedness measure for computing semantic relatedness of terms and texts are presented and benefits and drawbacks of different approaches to evaluating semantic relatedness are discussed. It must be noted that this survey will be a valuable reference to researchers of lexical semantic relatedness and can substantially support the research activities in this field.