BEE INSPIRED DETECTING AND TRACKING OF CURRENTLY DEVELOPING NEWS STORIES FROM THE WEB

摘要

In this paper we present an approach to detection and tracking of currently unveiling news stories through analysis of news articles obtained from the Web. Articles are obtained using algorithm inspired by honey bees foraging for food. Due to the distributed dynamic nature of communication protocol utilized among bee workers when foraging for food, we are able to detect, evaluate and retrieve articles related to current news events from the web. During this process key terms related to news stories are identified and relationships between articles are established based on these terms. This results in a graph of nodes modelling the article space. The graph is updated and appended in a flexible way based on the real time changes in the articles. News stories are subsequently extracted from the graph using modularity optimizing graph algorithms.