-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathRough.txt
181 lines (129 loc) · 5.85 KB
/
Rough.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
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
0000
0110
1001
0010
1010
0001
1000
0110
0000
0100
1000
0110
0100
1000
0010
0100
XORROR1 S = 0010
XORRO2 S = 0011
Response = 0
============================
2.9105443427999944 6.189227954799992 88156.0 0.9321249999999999
3.1133923138000683 7.429136546200016 87706.0 0.9197749999999999
1.4055762991999472 6.472391509600129 83876.0 0.9445
===========================
CS971
1.502902739995625 2.2300484200008213 88152.0 0.9321249999999999
CS983
4.084645380004076 2.193555680004647 84118.0 0.9378949999999999
=============
1.2363050096000052 6.360938097799999 83876.0 0.9445
1.3912564874000055 6.249904980399992 84716.0 0.947425
========
fit_intercept=True - decreases the accuracy
================
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=0.01)
2.0125108240000147 6.699569712799985 88996.0 0.825225
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=1)
2.2023023144000033 7.261396257000001 88156.0 0.9390750000000001
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=10)
2.846358982999982 6.852763996199974 88156.0 0.94945
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=11)
3.1303430156000105 7.039806393000072 88156.0 0.94955
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=20)
3.3007079377999617 6.654927636799949 88156.0 0.9496499999999999
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=40)
3.20083408240007 6.624912874600023 88156.0 0.949975
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=60)
3.729819408999947 6.229424281200044 88156.0 0.9496499999999999
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=100)
3.4016365134000353 6.707200896399991 88156.0 0.9492749999999999
===============
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=10, tol=1e-10)
4.239220157800037 6.45599754459995 88156.0 0.949475
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=11, tol=1e-10)
3.813806633999866 7.09931211159992 88156.0 0.94955
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=11, tol=1e-4)
2.130466009800057 7.491350201799923 88156.0 0.94955
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=11, tol=1e-1)
1.0446536763998666 7.026190121400032 88156.0 0.790625
============
clf=LogisticRegression(max_iter=10000, solver='newton-cg', C=11, tol=1e-4,penalty='l2')
3.14519290979988 6.486940754199895 88156.0 0.94955
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=10, tol=1e-4,penalty='l2')
1.3102126592000787 7.1384906877998215 88156.0 0.9484999999999999
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=11, tol=1e-4,penalty='l2')*****
0.9914572810000208 7.048575445200004 88156.0 0.9487500000000001
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=11, tol=1e-4,penalty='l1')
3.7356729968000764 6.412336455400146 88156.0 0.9476600000000002
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=11, tol=1e-10,penalty='l2')
1.1685585180000089 7.284882889199889 88156.0 0.9487
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=40, tol=1e-4,penalty='l2')
1.5100987823999277 6.74268217459994 88156.0 0.94945
clf=LogisticRegression(max_iter=10000, solver='liblinear', C=1, tol=1e-4,penalty='l2')
0.7836511866000364 7.345262839800034 88156.0 0.9403749999999998
=============
If you prioritize accuracy, go for C=40 with the newton-cg solver.
If you need a balance of speed and accuracy, consider C=11 with the liblinear solver for faster training while still achieving high accuracy.
Final Recommendation
Given your goal of avoiding overfitting while maintaining high accuracy, I recommend the following options:
For Highest Accuracy:
C=40, solver='newton-cg' (94.99% accuracy)
For Best Performance with Speed:
C=11, solver='liblinear', tol=1e-4, penalty='l2' (94.88% accuracy)
For Good Balance:
C=10, solver='newton-cg' (94.95% accuracy)
Conclusion
Choose C=40 for optimal accuracy, or consider the liblinear solver with C=11 for faster performance while still achieving high accuracy.
Monitor validation performance to ensure your choice generalizes well to unseen data.
===============
clf=LogisticRegression(max_iter=10000, C=0.01)
1.6619612469999994 7.036595888600004 88992.0 0.825175
clf=LogisticRegression(max_iter=10000, C=1)
2.1194421102000205 6.960793899999999 88152.0 0.939325
clf=LogisticRegression(max_iter=10000, C=10)
3.0364275532001104 6.494699841800047 88152.0 0.9498
clf=LogisticRegression(max_iter=10000, C=11)
2.8936114139999516 6.509436231000018 88152.0 0.94985
=================
clf=LogisticRegression(max_iter=10000, tol=1e-10)
3.0170075201999964 5.907416339600007 88992.0 0.939175
clf=LogisticRegression(max_iter=10000, tol=1e-3)
1.7349334075999878 7.106920074800018 88992.0 0.938775
clf=LogisticRegression(max_iter=10000, tol=1e-4)
2.558908140799974 6.835501674799991 88992.0 0.939325
clf=LogisticRegression(max_iter=10000, tol=1e-1)
1.435498751199998 6.876757038999972 88992.0 0.7798
clf=LogisticRegression(max_iter=10000, tol=1e1)
1.2304760762 7.390591038600019 88992.0 0.46299999999999997
==================
clf = LinearSVC(max_iter=10000, C=0.01)
1.0418885699999918 7.182633459000135 84716.0 0.903975
clf = LinearSVC(max_iter=10000, C=0.1)
0.9719954693999171 7.187399452400041 84716.0 0.9399999999999998
clf = LinearSVC(max_iter=10000, C=1)
0.9099672247998569 7.388745800799915 83876.0 0.947425
clf = LinearSVC(max_iter=10000, C=10)
1.2965315207999992 6.985495421199903 83876.0 0.9445
=================
clf = LinearSVC(max_iter=10000, tol=1e-10)
1.20513244699996 7.205967684199822 84716.0 0.947425
clf = LinearSVC(max_iter=10000, tol=1e-4)
0.9596972766000362 7.101524155999959 84716.0 0.947425
clf = LinearSVC(max_iter=10000, tol=1e1)
1.2211389825999504 6.716436382000029 84716.0 0.947425
clf = LinearSVC(max_iter=10000, tol=1e10)
1.373592323199955 7.140198136199979 84716.0 0.947425
=============
clf = LinearSVC(max_iter=10000, penalty='l2')
1.395721565399981 6.855036144000041 84716.0 0.947425