A Constructive and Hierarchical Self-Organising Model in a Non-Stationary Environment

Several related self-organising neural models have been proposed to enhance the flexibility of self-organising maps (SOM). These models are focused on the ability of continuous learning in a non-stationary environment. In our studies, these models depend on the pre-definition of several thresholds which are used as guidance of neural behaviours for specific data sets. However, it is not trivial to determine those thresholds in a nonstationary environment. When a proper threshold has been determined, this threshold may not be suitable for the future. Therefore, in this paper, we compare the dynamic adaptive selforganising hybrid (DASH) model with the growing neural gas (GNG) model by introducing several different initial thresholds to test their feasibility. Our experiments show that the DASH model is more stable and practicable for document clustering in a non-stationary environment since DASH adjusts its behaviour not only by modifying its parameters but also by an adaptive structure.

@InProceedings{HW05, 
 	 author =  {Hung, Chihli and Wermter, Stefan},  
 	 title = {A Constructive and Hierarchical Self-Organising Model in a Non-Stationary Environment}, 
 	 booktitle = {International Joint Conference in Neural Networks},
 	 editors = {},
 	 number = {},
 	 volume = {},
 	 pages = {2948--2953},
 	 year = {2005},
 	 month = {Jan},
 	 publisher = {IEEE},
 	 doi = {10.1109/IJCNN.2005.1556394}, 
 }