Database Reference
In-Depth Information
Figure 7.5
Entropy of coin flips, where X=1 represents heads
For the bank marketing scenario previously presented, the output variable is
subscribed
. The base entropy is defined as entropy of the output variable, that
is
.
As
seen
The next step is to identify the conditional entropy for each attribute. Given an
attribute , its value , its outcome
,
and its value , conditional entropy
is
the remaining entropy of given , formally defined as shown in
Equation 7.2
.
Consider the banking marketing scenario, if the attribute
contact
is chosen,
= {
cellular
,
telephone
,
unknown
}. The conditional entropy of
contact
considers all three values.
the table displays the probabilities of each value of the attribute. The next two rows
contain the probabilities of the class labels conditioned on the
contact
.