Unsupervised learning Clustering methods
E. Czogała, Jacek M. Łęski · Studies in fuzziness and soft computing · 2000
In the learning method described in the previous chapter we assume that we have target (desired) output of network for inputs from training data set. In contrast to that, in this chapter we use data set without the desired output of network. Such an approach to network learning without a teacher or supervisor is called an unsupervised method. The effect of that learning are features, regularities and structure of data extraction, and sometimes it is called a method that search for structures of data. For example in biology and medicine, where sets of physical and biochemical measurements define species and diseases, respectively, unsupervised methods are very useful. In this book unsupervised methods will be used to search fuzzy if-then rules. Grouping found by unsupervised methods is frequently referred to as clusters. The cluster is a natural and homogeneous subset of data. The data in each cluster are as similar as possible to each other, and as different (dissimilar) as possible from other cluster’s data. These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.