A MetriServe exclusive, our online reporting
tool includes a state-of-the-art web data mining algorithm that
looks for hidden patterns in the way visitors use your website.
How it Works
The MetriServe web data mining algorithm is based on a novel
sequential pattern mining technique first published at the 4th
Pacific-Asia Conference on Knowledge Discovery and Data Mining.
It is used to mine through all assembled web metrics pertaining
to web site navigation and to identify hidden relationships
between pages on your website even though they might not necessarily
be directly linked to each other. For example, if your website
is an online hardware store you might find that most visitors
browsing for hammers are also browsing for nails. You can use
this information to make the nails product page more easily
accessible from the page describing hammers and thereby increase
sales.
Configurable
The algorithm is executed in real-time and can be fine-tuned
with the following parameters:
Page: Indicates what page you are interested
in. The algorithm will return patterns related only to that
page. Alternatively all patterns found for all pages of
your web site will be reported.
Date Span: indicates what historical
time period has to be observed when looking for patterns.
Support: a value between 0.0 an 1.0 that
serves as a threshold. If you select a high support the
returned patterns will be fewer but will have stronger relations.
If you select a low support, more patterns will be returned
but many will have weaker relations.
