The distance matrix extension provides the following nodes:
Distance Matrix Reader
Read a distance matrix from an ASCII file. The file is supposed to contain the typically squared distance matrix, whereby it may optionally contain row and column headers (whose orders need to be equal). It also supports to read either the upper or lower triangular matrix (since distance matrices are typically symmetric).
Distance Matrix Writer
Write a distance matrix to an ASCII file; reverse operation to the Distance Matrix Reader node.
Distance Matrix Calculate
Compute a distance matrix from an input table and append the result as a new column to the table. The user has to specify the columns that are used for the distance calculation and the distance function (currently available: Euclidean, Manhattan and Tanimoto for bit vectors).
Run a k-Medoids clustering algorithm on a distance matrix. The k-medoids clustering technique is similar to the well k-means clustering, though it is not based on a numeric feature space and does not require a calculation of means.
New Hierarchical Clustering
Calculates a hierarchical clustering on a distance matrix. This node extends the hierarchical cluster node that is already available in KNIME by also accepting distance matrices as input. Furthermore it splits up the tasks of calculating the complete clustering (creation of the dendrogram) and the determination of the final clusters (cutting the dendrogram to output a given number of clusters).
Hierarchical Cluster Assigner
Uses the model learned by the New Hierarchical Clustering node and applies it to the given input data. It assigns the (training) data to the clusters, whereby the user has to specify the number of desired clusters in the dialog.
Hierarchical Cluster View
Displays the dendrogram plot of a hierarchical clustering.
Interactively collects hilite information during the node's execution. This can be very useful when the user manually browses the data and identifies interesting structures that he wants to annotate. This node is not related to Distance Matrices but is contained here because it was part of the same (sponsored) project.