Partitioning clustering decomposes a dataset into a set of disjoint clusters. Given a dataset, a partitioning method constructs several partitions of the data, with each partition representing a cluster. These methods relocate instances by moving them from one cluster to another, starting from an initial partitioning. Such methods typically require that the number of clusters be preset by the user. To achieve global optimality in partitioning-based clustering, an exhaustive enumeration process of all possible partitions is required. The following figure shows how a partitioning clustering constructs several partitions of the data:

In partitioning clustering, we select a set of parameters in advance. These parameters are then adjusted to optimally satisfy a chosen criterion of separation and compactness of our clusters. In this way, a hidden feature of the data can be highlighted, revealing the partitioning key. Typically, this leads to minimizing a measure of dissimilarity in the samples within each cluster, while maximizing the dissimilarity of different clusters. For example, we can use centers to represent clusters and then improve the partitioning by moving objects from group to group.