The congruity between linkage‐based factors and content‐based clusters—an experimental study using multiple document corpora

Abstract

Intellectual Structure ( IS ) is a bibliometric method that is widely applied in knowledge domain analysis and in science mapping. An intellectual structure consists of clusters of related documents ascribed to individual factors. Documents ascribed to a factor are generally associated with a common research theme. As such, the contents of documents ascribed to a factor are theorized to be similar to each other. This study shows that the link‐based relatedness implies content‐based similarity. The intellectual structures of two research domains were derived from data sets retrieved from the Microsoft Academic Search database. The collection of documents ascribed to a factor is referred to as a factor‐based document cluster, which the content‐based document clusters are compared with . All documents in an intellectual structure are re‐clustered based on their content similarity, which is derived from the cosine of their vector form encoded in documents' term frequency inverse document frequency ( TF‐IDF ) weighted terms. The factor‐based document clusters are then compared with the content‐based cluster s for congruity. We used the Rand index and kappa coefficient to check the congruity between the factor‐based and content‐based document clusters. The kappa coefficient indicates that there is fair to moderate agreement between the clusters derived from these two different bases.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.