Visualization of association graphs for assisting the interpretation of classifications

TL;DR

This paper presents visualization tools for association graphs derived from abstracts to aid interpretation of classifications, using interactive interfaces for exploring author and term relationships.

Contribution

It introduces novel visualization interfaces for association graphs, integrating co-author and author-term relationships with interactive exploration features.

Findings

Effective visualization of association graphs enhances interpretability.

Interactive interfaces facilitate exploration of complex bibliographic data.

System successfully maps and clusters author-term relationships.

Abstract

Given a query on the PASCAL database maintained by the INIST, we design user interfaces to visualize and browse two types of graphs extracted from abstracts: 1) the graph of all associations between authors (co-author graph), 2) the graph of strong associations between authors and terms automatically extracted from abstracts and grouped using linguistic variations. We adapt for this purpose the TermWatch system that comprises a term extractor, a relation identifier which yields the terminological network and a clustering module. The results are output on two interfaces: a graphic one mapping the clusters in a 2D space and a terminological hypertext network allowing the user to interactively explore results and return to source texts.