An efficient clustering method of the SOM based on genetic algorithm with feature weighting

The clustering result of SOM(Self-Organizing Maps) neural network is affected by feature weighting values of input data. This paper presents a SOM clustering method based on genetic algorithm. The genetic algorithm is utilized to search optimal feature weighting values through updating its fitness, and this updating process is realized by enlarging the distance of between-cluster and decreasing the distance between the winner neurons and the input data. This method can improve the clustering recognition rate of the SOM. Computer simulation confirms its validity.