K-means, SOM, k-nn or classical clustering methods?

The best-known optimization clustering algorithm is k-means clustering. Unlike hierarchical clustering methods that require processing time proportional to the square or cube of the number of observations, the time required by the k-means algorithm is proportional to the number of observations. This means that k-means clustering can be used on larger data sets. In fact, k-means clustering is inappropriate for small ( 100 observations) data sets. If the data set is small, the k-means solution becomes sensitive to the order in which the observations appear (the order effect).