cluster analysis

Discussion created by campgl01 on Nov 28, 2013
i am trying to defined landscape areas with the same characteristics  based on different variables. I have different rasters  mdt,  slope map, insolation map, distance to towns and the CORINE LAND COVER (land cover map) so i did a PCA followed by a cluster. To do the cluster i used the iso cluster unsupervised classification tool form spatial analysis tool--> multivariate. i tryed different nunber of classes but when i checked the classes they are wrong because  any difference, in terms of land cover, is done. When the variable land used is considered, the largest class in the original land cover map is always the one that wight the most in the cluster clases so no difference is made between an urban area or a forest.
Could someone help me with this problem? or can someone tell me how i could do the cluster to deffine this areas, might i  be using the wrong tool?