The ISODATA algorithm is a very classic unsupervised learning dynamic clustering algorithm based on statistical pattern recognition and has strong practicability. The ISODATA algorithm can not only complete the cluster analysis of samples by adjusting the categories to which the samples belong, but also automatically "merge" and "split" categories, thereby obtaining clustering results with a reasonable number of categories.
Expand
