KAnt: Leveraging ant colony optimization for automatic knowledge acquisition from web documents

摘要

This paper suggests a novel algorithm (KAnt) inspired by ant colony optimization strategies for knowledge acquisition. KAnt algorithm attempts to devise a unique solution for eminent knowledge acquisition problem of losing interest in content rich documents due to low familiarity. We utilize our solution to work with web based documents, considering documents as nodes in a graph problem. Locating content rich documents is achieved through intelligent ants that are equipped with numerical statistic for document identification. Documents are found via pheromones deposited by such ant colony. Experimental results acquired through domain expert evaluation show that our proposed approach has contributed for knowledge acquisition remarkably.