It may be also interestingto note that the basic idea of RPCL
can be regarded as a kind of unsupervised extension of Kohonen’s supervised learning vector quantization algorithm LVQ2
[15] which can give a result closer in effect to Bayes decision
theory by simultaneously modifying the weight vectors of both
the winner and its rival when the winner is in a wrong class
but the rival is in a correct class for an input vector.
In the sequel, the characteristics of RPCL will be further
illustrated through some experiments