TrAnET: Tracking and Analyzing the Evolution of Topics in Information Networks

Abstract

This paper presents a system for tracking and analyzing the evolution and transformation of topics in an information network. The system consists of four main modules for pre-processing, adaptive topic modeling, network creation and temporal network analysis. The core module is built upon an adaptive topic modeling algorithm adopting a sliding time window technique that enables the discovery of groundbreaking ideas as those topics that evolve rapidly in the network.