Information Technology Reference
In-Depth Information
Table 8.4 A data set based
on the lower approximation
{2} of the concept {1, 2}
Case Attributes
Decision
Temperature Headache
Nausea Flu
1
High
Ye s
No
SPECIAL
2
Very_high
No
No
Ye s
3
Very_high
Ye s
No
SPECIAL
4
Normal
No
No
SPECIAL
5
High
No
Ye s
SPECIAL
6
Normal
Ye s
Ye s
SPECIAL
7
High
Ye s
No
SPECIAL
Table 8.5 A data set based
on the upper approximation
{1, 2, 7} of the concept {1, 2}
Case Attributes
Decision
Temperature Headache
Nausea Flu
1
High
Ye s
No
Ye s
2
Very_high
No
No
Ye s
3
Very_high
Ye s
No
SPECIAL
4
Normal
No
No
SPECIAL
5
High
No
Ye s
SPECIAL
6
Normal
Ye s
Ye s
SPECIAL
7
High
Ye s
No
Ye s
Thus, if we are going to use the strategy of rule induction based on feature selec-
tion, possible rules induced from Table 8.3 are:
(
Temperature
,
high
)
and
(
Headache
,
yes
) (
Flu
,
yes
),
(
Temperature
,
very _ high
)
and
(
Headache
,
no
) (
Flu
,
yes
),
(
,
)
(
,
) (
,
),
Temperature
high
and
Headache
yes
Flu
no
(
,
)
(
,
) (
,
),
Temperature
very _ high
and
Headache
yes
Flu
no
(
Temperature
,
normal
)
and
(
Headache
,
no
) (
Flu
,
no
),
(
Temperature
,
high
)
and
(
Headache
,
no
) (
Flu
,
maybe
),
).
At the same time, the LEM2 algorithm will induce the flowing set of possible
rules from the same data set:
(
(
Temperature
,
normal
)
and
(
Headache
,
yes
) (
Flu
,
maybe
Nausea
,
no
)
and
(
Temperature
,
high
) (
Flu
,
yes
),
(
Temperature
,
very _ high
)
and
(
Headache
,
no
) (
Flu
,
yes
),
(
Nausea
,
no
)
and
(
Headache
,
yes
) (
Flu
,
no
),
(
Temperature
,
normal
)
and
(
Headache
,
no
) (
Flu
,
no
),
(
).
Again, it is quite clear that LEM2 produces simpler rules (and fewer rules).
Nausea
,
yes
) (
Flu
,
maybe
 
Search WWH ::




Custom Search