On Dynamic Clustering Models for 3-way Data
Mika Sato‐Ilic · Journal of Advanced Computational Intelligence and Intelligent Informatics · 1999
This paper presents clustering models for 3-way data consisting of similarities of objects over several times. Conventional methods for such 3-way data yields results showing the latent structure of inheritance over times, but cannot show exact changes of clusters over times. The advantage of proposed techniques for classifying 3-way data is showing exact changes, therefore the results for several times are comparable on the same coordinate of the solution’s space. The validity of these models is demonstrated by several examples.