Information Technology Reference
In-Depth Information
Fig. 4.10
Zoom in :::
Fig. 4.11
Details on demand
When the user moves his/her mouse over a document in the thematic space, the
title is flashed out on the screen. When the user clicks on the document, the content
of the document becomes available. When the user has decided that the current
document is relevant for the task, he/she can mark the document.
First, we use two users' trails as the training set to build the first Hidden Markov
model œstate. We choose user jbr and nol because one marked the most documents
and the other clicked the most number of times.
The third parameter of a Hidden Markov model is the intial distribution, denoted
as  . Intuitively, this is the likelihood that users will start with a given document for
their information foraging.
Search WWH ::




Custom Search