LVQ with a weighted objective function
Su-Jeong You, Chong‐Ho Choi · 2002
In competitive learning neural network, pattern clustering is one of the main research areas. Many competitive neural networks are based on vector quantization. Depending on the method of choosing the representative weight vectors, competitive neural networks have a great variety of algorithms. In this paper, an algorithm, a variety of GLVQ, is proposed and is compared with other algorithms. It is shown from simulation results that the proposed algorithm gives better performance than other algorithms in clustering.