Dustin Stevens-Baier
COMP 578
8-31-06
Assignment
#1
Suppose that you are employed as a data mining consultant for an
Internet search engine company. Describe how data mining can help the
company by giving specific examples of how techniques, such as clustering,
classification, association rule mining, and anomaly detection can be
applied.
Association rule
mining would be used for determining what web pages are accessed together, so
you can better direct traffic, or sell data to other companies so they no where
to market. For example if espn is accessed by the same people that access
sports illustrated then SI would know to advertise their product on espn.
They would also know to advertise anywhere else people who go to espn go.
This could also be applied to location if people from a certain location look at
specific sites then companies in that area know where to
advertise.
Clustering analysis can be used to group certain sites
together. This is extremely useful for a search engine because if I
am searching for espn it makes since that my result should be the different espn
site, like radio fantasy, classic and the normal site. I probably
shouldn't get pages that have to do with cooking or politics. The more
sites that I can get that are close together and related to each other the
better my search engine is and the more people will use it. This data
can also be sold so people know how to set up their site to be similar to
popular sites and get higher up on the search
match.
Anomaly detection is important because the data
that is used either for clustering or for association needs to be as accurate as
possible so that it is useful to our company so that we can accurately express
who, what, and where are searches are coming from. This is extremly
important if we are selling our data becuase we want to make sure that people
are getting accurate data so that we get repeat customers.