Optimized Computational Intelligence Driven Decision-Making

8Visualization of Textual Corpora Using Social Network Analysis

Indu Rodda* and Durga Bhavani S.

University of Hyderabad, Hyderabad, Telangana, India

Abstract

As more people lean toward movies and television shows, the practice of reading novels is gradually dwindling. This study uses social network analysis to make novel visualization more enjoyable. A character network is generated using the book’s main characters as nodes and the interactions between them as edges. Cumulative sub-graphs of the chapters in the novel are used to generate network snapshots. The main goal of the visualization is to depict how the main characters’ interactions change as the plot progresses by applying the community discovery for dynamic networks methodology. Additionally, using a similarity score, a mapping algorithm is developed to map the communities between two successive snapshots. This algorithm recognizes nearly invariant communities, community mergers, and the emergence of new communities. Additionally, the communities are given creative color assignments, so the viewer can easily detect community changes. The implementation is carried out for one of J.K. Rowling’s well-known novels, and the outcomes are displayed for chapters highlighting essential plot points. The implementation’s results demonstrate that the color scheme and similarity score both successfully depict the dynamics of community transitions.

Keywords: Dynamic networks, community discovery, similarity measure, visualization ...