Descriptive feature score

The manual says "...vcluster displays a number that in the case of the descriptive features is the percentage of the within cluster similarity that this particular feature can explain. For example, for the 0th cluster, the feature “warrior” explains 38.4% of the average similarity between the objects of the 0th cluster."

If I am using cosine based similarity scores, what exactly does the 38% measure? What is the model for "explains"?


RE: What it means is that if you

What it means is that if you set that feature to 0, then the similarity will drop by about 38%.
So if it was .25, now it will be ~.16.