Latest additions (August 7, 2012):
- Incorporated in-bound links into relevancy ranking criteria.
The new algorithm counts all in-bound links for all pages and stores this as a "link count" for each indexed document.
This requires two passes, one to count all of the links and create a link file, and then a second pass to index
the documents with the in-bound link counts. The CloudSearch rank expressions were then re-calculated to incorporate
this new statistic.
Try it! The results are even better than before!
Bookmark this page, this demo will continue to improve throughout 2012.
This site is a laboratory for experimenting with innovative new ways to search.
- We are using Wikipedia as our demonstration data set because it's large, well known, and, well, freakin' awesome
- Amazon CloudSearch is providing the search results; a great choice when you donít know how much search traffic you are going to get, don't want to have to worry about uptime, but you do want access to a wide range of cool search engine functions
- Aspire is providing content processing prior to indexing into Amazon CloudSearch. The role that Aspire plays is described here.
- To learn how this search experience was created see our blog.