Semantic relationships between highly cited articles and citing articles in information retrieval

A major concern in using citation data for content analysis resides in the claim that the content analysis and indexing of texts are improved by using citation data in addition to terms extracted from the original texts. This claim is based on the assumption that documents with citation relationships are semantically related. By investigating this assumption, it is our hope to uncover why the previous studies on using citation data for document indexing are controversial. In this study, it is indicated that there is a statistically significant semantic relationship between highly cited publications and their citing publications in comparison to highly cited publications with no citing relationship, within the same topic literature. Visualizing this set of data, however, shows that citation-based data represents different types of subject relevance.