Semantic Clustering in Dutch

1 Choose k cluster centers, which are usually k randomly-chosen patterns or k randomly defined points inside the vector space. 2 Assign each pattern to the closest cluster center (using the cosine measure). 3 Recompute the cluster centers using the current cluster memberships. 4 If a convergence criterion is met (e.g. no reassignment of patterns to new cluster centers), stop the algorithm. Otherwise, go to step 2. 1 Take each individual pattern in the pattern set to form a cluster. 2 The two clusters which are most similar are grouped together. Most similar means: the two clusters with the smallest distance between the averages of the clusters. 3 Step two is repeated until there is only one cluster left. When the algorithm terminates, all clusters are hierarchically connected to the root node.