MRMD: Maximum-Relevance-Maximum-Distance
|
|
|
|
(1) single-lable (arff format)
Assume that the input file has a total of 5 examples, 4 features and 1 class standard:
@relation test
@attribute Fea1 numeric
@attribute Fea2 numeric
@attribute Fea3 numeric
@attribute Fea4 numeric
@attribute label {1,0}
@data
12,25,26,31,1
25,26,35,95,1
56,96,54,85,0
54,63,69,58,0
56,58,23,21,1
Each line is an example with four features and the last one is the label.
(2) multi-lable (mulan arff format)
Assume that the input file has a total of 5 examples, 4 features and 3 class standard:
@relation test
@attribute Fea1 numeric
@attribute Fea2 numeric
@attribute Fea3 numeric
@attribute Fea4 numeric
@attribute label1 {1,0}
@attribute label2 {1,0}
@attribute label3 {1,0}
@data
12,25,26,31,1,0,1
25,26,35,95,1,0,1
56,96,54,85,0,1,0
54,63,69,58,0,1,1
56,58,23,21,1,1,0
Each line is an example with four features and the last three is the label.
(3)multi-lable (MEKA arff format)
Assume that the input file has a total of 5 examples, 4 features and 2 class standard:
@relation test
@attribute label1 {1,0}
@attribute label1 {1,0}
@attribute Fea1 numeric
@attribute Fea2 numeric
@attribute Fea3 numeric
@attribute Fea4 numeric
@data
1,0,12,25,26,31
1,0,25,26,35,95
0,1,56,96,54,85
0,1,54,63,69,58
1,1,56,58,23,21
Each line is an example with four features and the last two is the label.
All Rights Reserved Copyright @ 2016|Dr. GuoJisheng
Last Modified in 2016/2/23