Sat Dec 29 00:04:00 PST 2018

179
Speculation
I mention a number of algorithms and concepts in the following section, including:
Hilltop, TrustRank, Topic-Sensitive PageRank, temporal analysis, and latent
semantic indexing (LSI
.
Some of these algorithms may not be part of the current search environment, but
the ideas contained within them are still worth understanding to see where search
may be headed and what search topics search engineers think are important to
improve their overall relevancy scores.
Local Re-ranking Results Based on Inter-Connectivity
Hilltop
Hilltop was an algorithm that reorganizes search results based on an expert rating
system.
In the Hilltop white paper, they talk about how expert documents can be used to
help compute relevancy. An expert document is a non-affiliated page that links to
many related resources. If page A is related to page B and page B is related to page
C, then a connection between A and C is assumed.
Additionally, Hilltop states that it strongly considers page title and page headings in
relevancy scores; in fact, these elements can be considered as important as, or more
important than, link text. It is likely that Hilltop also considers the links pointing
into the page and site that your links come from.
The benefit of Hilltop over raw PageRank (Google
is that it is topic sensitive, and
is thus generally harder to manipulate than buying some random high-power off-
topic link. The benefits of Hilltop over topic distillation (the algorithm that powers
Ask.com, which will be discussed later
are that Hilltop is quicker and cheaper to
calculate and that it tends to have more broad coverage.
When Hilltop does not have enough expert sites, the feature can be turned off, and
results can be organized using a global popularity score, such as PageRank.
Google might be using Hilltop to help sort the relevancy for some of their search
results, but I also see some fairly competitive search queries where three of my sites
rank in the top eight results. On those three sites, it would be fairly obvious for
search engines to know that they were all owned by me.
They may use something like Hilltop to scrub the value of some nepotistic links,
but it will not wipe out all related sites just because they are related. When you
search for things like Microsoft, it makes sense that many of the most relevant
websites are owned by the same company.