meaningful and stable vs. not
On Web4Lib a while ago we were discussing clustering vs faceted vs federated search (all very different!). This reminded me of work by an HCIL-er
(now professor at CUA
)... I've made this point here before, but little has changed.
Meaningful and stable (yes, not the right database for the search, probably):
Facets - author, affiliation, publisher, language, year, classification code, controlled vocabulary...
Suggestions from the CV (not really suggestions but frequently occurring descriptors):
Not meaningful, maybe stable:
Ok, so here are the clusters for "nerve blockade"
- Chi-Square Test
- Prospective Studies
- Middle Age
- Male Female (hopefully one or the other or both will be involved)
- Random Assignment
- Continuing Credit
- Comparative Studies
- Descriptive Statistics
Labels: clustering, faceted, research databases