Create clusters of similar items

Generate clusters of similar items based on a clustering threshold, which is the minimum level of matching for items to be considered similar to each other.

Prerequisite:

• You de-duplicated the archive.

1. On the Archive Details page, in the Clustering section, click Generate Clusters.
2. From the Clustering Threshold list, select the minimum matching threshold.

   If you set a lower threshold, the clusters are less accurate. However, if you set a higher threshold, more clusters are created.

3. Click Perform Clustering.

   Note: The processing time depends on the number of items in the archive.

4. Optional: After the files have been processed, you can do the following from the Archive Details page:

• To view the list of clusters, click View clusters.

You can rename a cluster or create a bucket or dataset from it.

• Click Sample

• Click one of the following:

                  - Add to new bucket, to create a bucket from the clusters.

                  - Add to existing bucket.

                  - Add to new dataset, to create datasets from the clusters.

                  - Delete clusters, to delete the clusters, so you can create new ones based on a

                    different clustering threshold.