Ranking Algorithms for Complex Networks

Ranking algorithms have been widely used in search engines like Google to return the most relevant and informative pages to the users. The ranking algorithms have also been utilized to characterize pages in the web network in an attempt to cluster them based on the similarity in their characteristics. However, as formulations of the algorithms depend on the link addition process, the same algorithms cannot be extended to other type of complex networks such as trading networks and online auction networks. In this monograph, we study the link addition process in trading networks, and point out the fundamental differences between web network and trading networks. Based on these differences, we formulate a family of new ranking algorithms for trading networks, and then extend them to be used for clustering purpose. Our proposed algorithms seem to be working excellently in real data sets and also are extendable as clustering methods.

Autorentext

Andri Mirzal receives PhD in Information Science and Technology from Hokkaido University and B.Eng in Electrical Engineering from Bandung Institute of Technology. His research interests are in Machine Learning and Web Search Engine.