Different approaches to automated classification : is there an exchange of ideas?
Koraljka Golub, Birger Larsen · 2005
Abstract. Automated classification of text has been studied by three major research communities, machine learning, information retrieval, and library science, each taking a different approach. The paper aims to study to what a degree the three communities explore others ’ ideas, methods, findings. To that purpose we studied direct links (do authors from one community cite authors from another) and indirect links (using bibliographic coupling). Although the study is based on a small sample of 148 papers, the results indicate that the three communities do not exchange ideas to a great extent.