-
Notifications
You must be signed in to change notification settings - Fork 1
/
ClassResults.txt
123 lines (95 loc) · 4.08 KB
/
ClassResults.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
#Entries of what I think about certain classifiers and how they are doing. I think we'll have to use filtered classifier for all anyways.
Zero R is dumb
~Bagging is awesome when used under voting! (Not suggested individually)
~VotedPerceptron is not very accurate (about 67%) but much better at clssifying for no's.
nvm, it's terrible when combined with others
~Random Commitee under meta is also not bad
~Random SupSpace works too
~J48 remains the best
~AdaBoost might work and you can add one more classifier to it
~SMO is good
~LogisticR under functions: 71.13%
~SGD under functions: 74.80%
~SMO under functions: 75.07%
~AttributeSelectedClassifier under meta: 76.12% (and you can add another classifier to it)
~Bagging under meta: 77.69% (and you can add another classifier to it)
~LogicBoost under meta: 77.95% (nvm DONT't use it)
~RandomSubSpace under meta: 76.38% (72.63 in 50-50 percent split. Better for no's!)
~Decision Table under rules: 75.59% (JRip about the same)
~PART under rules: 76.64%
~NaiveBaysMultinomialText: 81.65%
(most bayes are quite good other than NaiveBayes and NaiveBayesUpdateable)
~SGD under functions: 81.84%
*Used Vote under meta and a combination of Random Forest and REP Tree:
Cross Validation: 77.39%
50-50 Percetage Split: 70.93%
Supplied Test for 12: 66.67%
*Used PART under rules:
Cross Validation: 75.07%
50-50 Percetage Split: 74.42%
Supplied Test for 12: 75%
(Slightly better at classifying for no's)
*Used Vote and a combination of Random Forest and PART:
Cross Validation: 75.36%
50-50 Percetage Split: 75.58%
Supplied Test for 12: 75%
*Used JRip under rules:
Cross Validation: 76.52%
50-50 Percetage Split: 75.58%
Supplied Test for 12: 66.67%
*Used Vote combination of JRip, Bagging, RandomForest, PART, and VotedPerceptron:
Cross Validation: 77.97%
50-50 Percetage Split: 75%
Supplied Test for 12: 58.33%
(Used Majority Voting)
*Used Vote combination of RandomForest, PART, and Bagging(with J48):
Cross Validation: 79.42%
50-50 Percetage Split: 73.26%
Supplied Test for 12: 75%
(Used Majority Voting)
*Used Vote combination of RandomForest, PART, and Bagging(with RandomForest):
Cross Validation: 80.58%
50-50 Percetage Split: 76.16%
Supplied Test for 12: 58.33%
(Used Majority Voting)
(Add another Bagging with RandomTree for even better results)
(Add also add a Bagging with PART section)
*Used Vote combination of PART, Bagging(with PART), and Bagging(with RandomForest):
Cross Validation: 80.05%
50-50 Percetage Split: 72.63%
*Used Vote combination of PART, Bagging(with PART), Bagging(with RandomForest), and RandomCommittee:
Cross Validation: 79.53%
50-50 Percetage Split: 75.26%
*Used Bagging(with PART):
Cross Validation: 80.05%
50-50 Percetage Split: 74.62%
*(WRITER6) Used Vote combination of RandomForest, PART, and Bagging(with RandomForest):
Cross Validation: 84.12%
50-50 Percetage Split: 89.41%
(Used Majority Voting)
(Add another Bagging with RandomTree for even better results)
(Add also add a Bagging with PART section)
(Add J48 to get 90.59% on 50 percent split, but 83.72% on cross-validation)
*(WRITER6) Used Vote combination of RandomForest, PART, and Bagging(with RandomForest):
Cross Validation: 84.12%
50-50 Percetage Split: 89.41%
(Used Majority Voting)
(Add another Bagging with RandomTree for even better results)
(Add also add a Bagging with PART section)
(Add another Bagging with J48 to get 88.53% on 50 percent split, but 84.72% on cross-validation)
*Used Vote combination of Bagging(with PART) and Bagging(with RandomForest):
Cross Validation: 86.60%
50-50 Percetage Split: 87.10%
(Added Bagging with J48 to get 86.83% on Cross Validation)
*Used Vote combination of Bagging(with PART) and Bagging(J48):
Cross Validation: 88.25%
50-50 Percetage Split: 87.75%
*Used Vote combination of Bagging(with PART) and Bagging(J48), and Bagging(BayesNet):
Cross Validation: 87.61%
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ON~NEW~DATA~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*Baggings (with RandomForest, RandomTree, J48, PART)
Cross-Validation: 82.88
Percent Split: 83.92
*Baggings (with RandomForst and PART)
Cross-Validation: ~85.15%
Percent Split: ~83.93%