Allow the user to choose the granularity of the clusters

[mikekucera]

Allow the user to choose the granularity of the clusters. For example, do they want a lot of clusters or just a few.

• AA supports several clusterMaker algorithms. We would need to figure out which parameters for each algorithm causes more/fewer clusters.
• Idea: Perhaps run the clustering algorithm with different granularities in the background and let the user choose which results to use based on the number of clusters it will produce.
• This might work better if we use MCODE for clustering because its quite fast.
• Crazy Idea: Remove the dependency on clusterMaker, use MCODE instead.

[mikekucera]

another idea: use k-means and let the user choose the value of k

[mikekucera]

Using MCL with the inflation parameter