Dear Sir. Concerning your issue about the best validation method for som clustering. The Self-Organizing Map (SOM) is a type of neural network suitable for unsupervised learning (Kohonen, 1997). SOMs combine competitive learning with dimensionality reduction by smoothing the clusters with respect to an a priori grid. One of the main characteristics of these networks is the topological ordering property of the clusters generated. Clusters objects are mapped in neighbor regions of the grid, delivering an intuitive visual representation of the clustering. SOMs are reported to be robust and accurate with noisy data (Mangiameli et al., 1996). On the other hand, SOM suffers from the same problems such as those of dynamical clustering: sensibility to the initial parameters settings and the possibility of getting trapped in local minimum solutions (Jain et al., 1999). The SOM method works as follows. Initially, one has to choose the topology of the map. All the nodes are linked to the input nodes by weighted edges. The weights are first set at random, and then iteratively adjusted. Each iteration involves randomly selecting an object x and moving the closest node (and its neighborhood) in the direction of x. The closest node is obtained by measuring the Euclidean distance or the dot product between the object x and the weights of all nodes in the map. The neighborhood to be adjusted is defined by a neighborhood function, which decreases over time. Evaluation of clustering results (or cluster validation) is an important and necessary step in cluster analysis, but it is often time-consuming and complicated work. We present a visual cluster validation tool, the Cluster Validity Analysis Platform (CVAP), to facilitate cluster validation. The CVAP provides necessary methods (e.g., many validity indices, several clustering algorithms and procedures) and an analysis environment for clustering, evaluation of clustering results, estimation of the number of clusters, and performance comparison among different clustering algorithms. It can help users accomplish their clustering tasks faster and easier and help achieve good clustering quality when there is little prior knowledge about the cluster structure of a data set. I think the following below links and the attached files may help you in your analysis: