An Incremental Neural Network for Non-stationary Unsupervised Learning

Shen Furao,Osamu Hasegawa
DOI: https://doi.org/10.1007/978-3-540-30499-9_98
2004-01-01
Abstract:A new online learning mechanism is proposed for unsupervised classification and topology representation. It can represent the topological structure of unsupervised online non-stationary data, report the reasonable number of clusters, eliminate noise, and give typical prototype patterns of every cluster without priori conditions such as a suitable number of nodes or a good initial codebook.
What problem does this paper attempt to address?