sarchitect

Eigen-value Clustering

Eigen value clustering is based on the principle that eigen vectors of the similarity matrix associated with the given data contain information on how the data cluster. 

Select Explore|Cluster|Eigen Value to open its window.

Parameters 

Cluster On: You can choose to cluster on rows, columns or both rows and columns.

Distance Metric: Select the distance metric to be used for calculating the distance between molecules and averages of all clusters. The options are Angular, Euclidean, Squared Euclidean, Manhattan, Chebychev, Differential, Pearson Absolute and Pearson Centered.  See Theory - Statistics for details on these measures.

Cutoff Ratio: This defines a cut off for isolating the cluster which rises to the top. A larger value imposes a more aggressive cutoff. A value of 0 would give just one large cluster, and the number of clusters increases as this cutoff is increased. The default is 0.9.

Select the data columns for the analysis in the Columns tab by selecting the columns and moving them to the Columns Selected pane.

Click OK to run. The results are output in a child dataset called EigenVector Clustering and consist of a cluster set, a dendrogram and a similarity image. For a description of these plots and extracting information from them, see Clustering Results.

See Also
Theory - Statistics

Sarchitect Help Table of Contents