The paper introduces a new clustering algorithm, called CLUNET, based on the neural network philosophy. The paper actually presents a neural network able to produce an unsupervised classification. The geometrical background is the Hamming space of the weights associated to the nodes obtained by a suitable transformation of the data in a multidimensional space. The space and the computational complexity is quite linear and the learning phase is very fast and really competitive with other clustering algorithms based both on the usual techniques and on neural networks philosophy.
Bellacicco, A., Tulli, V. (1998). A neutral network based clustering algorithm: CLUNET. METRON, LVI(3-4), 173-184.
A neutral network based clustering algorithm: CLUNET
Tulli,V
1998
Abstract
The paper introduces a new clustering algorithm, called CLUNET, based on the neural network philosophy. The paper actually presents a neural network able to produce an unsupervised classification. The geometrical background is the Hamming space of the weights associated to the nodes obtained by a suitable transformation of the data in a multidimensional space. The space and the computational complexity is quite linear and the learning phase is very fast and really competitive with other clustering algorithms based both on the usual techniques and on neural networks philosophy.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.