|
ECIS automatically and dynamically groups search results into categories that are intelligently created from the words and phrases contained in them. By clustering similar results into multiple categories and labelling each category with a representative title, the system allows
you to explore and analyze the results by concepts.
Default clustering strategy
ECIS categorizes the search results based on the content of four attributes: title, abstract, body and collection. Thus, new categories are created on-the-fly from the actual data. New terms emerging within a given domain
are automatically picked up by the system as their occurrence increases.
|
 |
|
Connection to high-quality corporate taxonomy is also supported using the collection attribute. Search results that are pre-categorized according to the taxonomy
are grouped together, as well as with similar new results using the value of collection.
Custom clustering strategy
You can create concepts around attributes to obtain a rapid and powerful analysis. Turn
ECIS into an expert locator application by simply clustering all search results on the author attribute. The system
is intelligent enough so has to group co-authors into a single category; or to generate sub-categories representing areas of interest for a very
specific author. Turn ECIS into a trend analysis application by clustering on the date. The system groups results by common year, common months or common weeks within a month.
To change the clustering strategy, follow the procedure below:
- Go to the main search page
- Click Cluster by at the top of the cluster area
A new window is displayed
- Click the name of the attribute to use the clustering
- After the clusters have been reorganized, the window closes
To revert to the default clustering strategy, follow the procedure below:
- Go to the main search page
- Click Cluster by, at the top
of the cluster area
A new window is displayed
- Click Topics
- After the clusters have been re-organized, the window closes
|
