African Wordnet as a tool to identify semantic relatedness and semantic similarity

Abstract

Semantic relatedness and semantic similarity play a vital role in information retrieval of words in natural language processing applications. The multilingual nature of South Africa affords people a superficial (as opposed to an expert) knowledge of many languages, resulting in them finding it difficult to distinguish between semantic relatedness and semantic similarity. Computing semantic relatedness and semantic similarity using African Wordnet could help in developing a proper understanding of the meaning of certain words in the various languages. This article explores how African Wordnet can be used to identify semantic relatedness and semantic similarity. The focus of this article will be on isiZulu as one of the selected languages used in the African Wordnet Project. The purpose of the African Wordnet Project is the development of aligned wordnets for African languages spoken in South Africa. This article further discusses and analyses the examples of semantic relatedness and semantic similarity in the chosen language with a view to giving accounts of the two terms by providing examples and the usages of the words.