Database Reference
In-Depth Information
model
M
K−
1
(the model that is based on
D
K−
1
) on the database
D
K−
1
is
significantly different from the validation error rate of the model
M
K−
1
over
d
K
.
.
Inconsistency of the DM algorithm
. The basic assumption is that the
DM
algorithms in use were proven to be consistent and therefore are not
prone to be a major cause for a change in the model. Although, this option
should not be omitted, this work does not intend to deal with consistency
or inconsistency measures of existing
d
K
consists of the accumulated set of records obtained in period
K
DM
algorithms. We just point out
here that the
algorithm that was chosen for our experiments (IFN)
was shown to be stable in our previous work [21].
As noted in sub-section 2.1, the data mining model is generally described
DM
ˆ
M
G
:
A ¬
T
A
T
as:
, that is a relationship between
and
. The first cause is
explained via a change in the set of attributes
,whichcanalsocausea
change in the target variable. For example, if the percent of women will rise
from 50% to 80% and women drink more white wine than men, than the
overall percent of white wine consumption (
A
) will also increase. This can
also affect the overall error rate of the data mining model if most rules were
generated for men. Also, if the cause of a change in the relevant period is
a change in the target variable (e.g., France has stopped producing white
wine, and the percent of white wine consumption has dropped from 50%
to 30%), it is possible that the rules relating the relevant attributes to the
target variable(s) will be affected.
Since it is not trivial to identify the dominant cause of change, this
work claims that a variety of possible changes might occur in a relevant
T
Table 1. Definition of the variety of changes in a data mining model.
Rules
AT
Description
−−−
No change.
−−
+ A change in the target variable.
−
+
−
A change in the attribute variable(s).
−
+ + A change in the target and in the input variable(s).
+
−−
A change in “patterns” (rules) of the data mining
model.
+
−
+ A change in “patterns” (rules) of the data mining
model, and a change in the target variable.
++
−
A change in “patterns” (rules) of the data mining
model, and a change in the input variable(s).
+
+ + A change in “patterns” (rules) of the data mining
model, and a change in the target and the input
variables.
