Thus, in this infinite family of collections, the Mutual Reinforcement approach is affected by the TKC effect (its ranking is biased in favor of tightly knit communities).
We note here that a special case of the TKC effect has been identified by Bharat and Henzinger .
The results for this query, with our first example of the TKC effect, are shown in Table II.
This query demonstrates the TKC effect on the WWW in a most striking fashion.
Again, we see the TKC effect: the Mutual Reinforcement approach ranks highly authorities on only one aspect of the query, while SALSA blends authorities from both aspects into its principal community.
We have also presented the TKC effect, a topological phenomenon which sometimes derails the MR approach and prevents it from converging to a useful community of authoritative sites.
The combination of a page's intracommunity authority score and its community's size is one of the factors that enable SALSA to blend authorities from different aspects of a multitopic query, and which reduces its vulnerability to the TKC effect.
These tests, as well as analytical consideration, have revealed a topological phenomenon on the Web called the TKC effect.
--In single-topic collections, the TKC effect sometimes results in the Mutual Reinforcement approach ranking many irrelevant pages as authorities.
We note that SALSA is less vulnerable to the TKC effect, and produces good results in cases where the Mutual Reinforcement approach fails to do so.