A COMPARISON OF SOM NEURAL NETWORKS AND K-MEANS CLUSTERING USING REAL WORLD DATA: CHINESE CONSUMER ATTITUDES TOWARDS IMPORTED FRUIT

X. Sun, R. Collins, J. Kim
SOM neural networks are regarded as a new clustering technique in market research. However for the technique to be widely adopted in practice, it should demonstrate superiority over traditional clustering methods. In this research we compared SOM neural networks with K-means algorithms to test their relative ability to generate reliable clustering solutions using real world data - Chinese consumer attitudes towards imported fruit. Results show that K-means performs better than SOM in terms of reliability, but SOM’s strength is its ability to discover the “natural” number of clusters.
Sun, X., Collins, R. and Kim, J. (2001). A COMPARISON OF SOM NEURAL NETWORKS AND K-MEANS CLUSTERING USING REAL WORLD DATA: CHINESE CONSUMER ATTITUDES TOWARDS IMPORTED FRUIT. Acta Hortic. 566, 185-191
DOI: 10.17660/ActaHortic.2001.566.21
https://doi.org/10.17660/ActaHortic.2001.566.21
Market segmentation, discriminant analysis, reliability tests, consumer behavior, clustering
English

Acta Horticulturae