Stable Deterministic Crystallization for Discovering Hidden Hubs
Yoshiharu Maeno, Yukio Ohsawa · 2006
Experts of chance discovery have recognized a new class of problems where the previous methods fail to reveal a latent structure behind observation. There are invisible events which play an important role in the dynamics of visible events. A hidden hub person (an invisible leader) in a communication network is a typical example. This paper presents a stable deterministic crystallization algorithm for discovering such hidden hub events. The algorithm is evaluated with the test data generated from a large scale-free random network. It is demonstrated that precision for discovering the hidden hub events remains as high as 80% to 100%, regardless of the prior knowledge and the network structure.