WorldCIST'15 - 3rd World Conference on Information Systems and Technologies

Full Program »

Ubiquitous Self-Organizing Map: Learning Concept-Drifting Data Streams

The Internet-of-Things promises a continuous flow of data where
traditional database and data-mining methods can not be applied. This
paper presents a novel variant of the well-known Self-Organized Map (SOM),
called Ubiquitous SOM (UbiSOM), tailored for streaming environments. This
approach allows ambient intelligence solutions that use exploratory
multidimensional clustering over a continuous data stream to provide
continuous exploratory data analysis. The average quantization error over
time is used for estimating the learn parameters, allowing the model to
retain an indefinite plasticity and to cope with concept drift within
a multidimensional stream.

Our experiments show that UbiSOM outperforms other SOM proposals in
continuously modeling concept-drifting data streams, converging faster
to stable models when the underlying distribution is stationary and
reacting accordingly to the nature of the concept-drift in continuous
real world data-streams.

Author(s):

Bruno Silva    
DI-FCT/UNL
Portugal

Nuno Cavalheiro Marques    
DI-FCT/UNL
Portugal

 

Powered by OpenConf®
Copyright ©2002-2013 Zakon Group LLC