### Abstract

We present a multi-layer competitive model to classify unknown data into hierarchical clusters through unsupervised learning. The first layer of the model, which is a Winner-Takes-All type competitive network within N reference vectors, is an extended Learning Vector Quantization whose distance measure is given by discriminant function of the Bayes decision. The extended LVQ is referred to Learning Baysian Quantization (LBQ). LBQ organizes reference vectors which correspond to prototypes of input data through unsupervised learning. The second layer of the model includes N LBQs as its subnetworks. The a-th reference vector in the first layer connects to the a-th LBQ in the second layer. The a-th LBQ works as a secondary LBQ if and only if the a-th neuron excites. When input data consists of several clusters which include several sub-clusters, the presented model organizes the prototypes of the clusters as the reference vector in the first layer and those of the subclusters as the reference vector in the second layer hierarchically.

Original language | English |
---|---|

Title of host publication | IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings |

Publisher | IEEE |

Pages | 815-818 |

Number of pages | 4 |

Volume | 2 |

Publication status | Published - 1997 |

Event | Proceedings of the 1997 6th IEEE Pacific Rim Conference on Communications, Computers and Signal Processing. Part 1 (of 2) - Victoria, Can Duration: 1997 Aug 20 → 1997 Aug 22 |

### Other

Other | Proceedings of the 1997 6th IEEE Pacific Rim Conference on Communications, Computers and Signal Processing. Part 1 (of 2) |
---|---|

City | Victoria, Can |

Period | 97/8/20 → 97/8/22 |

### Fingerprint

### ASJC Scopus subject areas

- Signal Processing

### Cite this

*IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings*(Vol. 2, pp. 815-818). IEEE.

**Competitive model to classify unknown data into hierarchical clusters through unsupervised learning.** / Meki, Yoshihisa; Kindo, Toshiki; Kurokawa, Hiroaki; Sasase, Iwao.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

*IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings.*vol. 2, IEEE, pp. 815-818, Proceedings of the 1997 6th IEEE Pacific Rim Conference on Communications, Computers and Signal Processing. Part 1 (of 2), Victoria, Can, 97/8/20.

}

TY - GEN

T1 - Competitive model to classify unknown data into hierarchical clusters through unsupervised learning

AU - Meki, Yoshihisa

AU - Kindo, Toshiki

AU - Kurokawa, Hiroaki

AU - Sasase, Iwao

PY - 1997

Y1 - 1997

N2 - We present a multi-layer competitive model to classify unknown data into hierarchical clusters through unsupervised learning. The first layer of the model, which is a Winner-Takes-All type competitive network within N reference vectors, is an extended Learning Vector Quantization whose distance measure is given by discriminant function of the Bayes decision. The extended LVQ is referred to Learning Baysian Quantization (LBQ). LBQ organizes reference vectors which correspond to prototypes of input data through unsupervised learning. The second layer of the model includes N LBQs as its subnetworks. The a-th reference vector in the first layer connects to the a-th LBQ in the second layer. The a-th LBQ works as a secondary LBQ if and only if the a-th neuron excites. When input data consists of several clusters which include several sub-clusters, the presented model organizes the prototypes of the clusters as the reference vector in the first layer and those of the subclusters as the reference vector in the second layer hierarchically.

AB - We present a multi-layer competitive model to classify unknown data into hierarchical clusters through unsupervised learning. The first layer of the model, which is a Winner-Takes-All type competitive network within N reference vectors, is an extended Learning Vector Quantization whose distance measure is given by discriminant function of the Bayes decision. The extended LVQ is referred to Learning Baysian Quantization (LBQ). LBQ organizes reference vectors which correspond to prototypes of input data through unsupervised learning. The second layer of the model includes N LBQs as its subnetworks. The a-th reference vector in the first layer connects to the a-th LBQ in the second layer. The a-th LBQ works as a secondary LBQ if and only if the a-th neuron excites. When input data consists of several clusters which include several sub-clusters, the presented model organizes the prototypes of the clusters as the reference vector in the first layer and those of the subclusters as the reference vector in the second layer hierarchically.

UR - http://www.scopus.com/inward/record.url?scp=0031369917&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031369917&partnerID=8YFLogxK

M3 - Conference contribution

VL - 2

SP - 815

EP - 818

BT - IEEE Pacific RIM Conference on Communications, Computers, and Signal Processing - Proceedings

PB - IEEE

ER -