Shen et al. propose an enhancement of a learning algorithm they developed previously. The base algorithm may be seen as a clever combination of self-organizing maps and neural gas. Since this algorithm is fairly new, literature sources related to it are rather scarce. I would recommend that anyone interested in learning more should consult the PhD thesis of the first author, which is available electronically [1].

The true value of this paper is in providing a means for online learning, which can only fully be appreciated after acquiring some background on the base algorithm through other papers or the above-mentioned dissertation. For static pattern classification, the algorithm in this paper is reported to be marginally better than support vector machines (SVMs) or growing neural gas (GNG), its main competitors. However, it is difficult to validate these claims, since reproducible demonstrations are not available.