An Incremental Neural Network for Non-stationary Unsupervised Learning

Shen Furao, Osamu HasegawaICONIP 2004 · OpenReview

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.
BibTeX
@inproceedings{furao2004incremental,
  author = {Shen Furao and Osamu Hasegawa},
  title = {{An Incremental Neural Network for Non-stationary Unsupervised Learning}},
  booktitle = {ICONIP},
  year = {2004},
  url = {https://openreview.net/forum?id=1CBTIwnT8tV},
}

Loading related content...

Figure