-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathKnnReal.py
176 lines (149 loc) · 6.28 KB
/
KnnReal.py
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
from collections import OrderedDict
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from pandas import DataFrame
from pandas import Series
import plotnine as gg
import sklearn as sk
import sklearn.model_selection as model_selection
from sklearn.model_selection import ShuffleSplit
import sklearn.feature_selection as feature_selection
import sklearn.neighbors as neighbors
import sklearn.pipeline as pipeline
import warnings
warnings.filterwarnings("ignore")
import pcaextractor
import MaclearnUtilities
from MaclearnUtilities import safeFactorize
plt.ion()
import RestrictedData
xs = RestrictedData.xs
xnorms = RestrictedData.xnorms
annots = RestrictedData.annots
ys = RestrictedData.ys
ynums = RestrictedData.ynums
def pandaize(f):
def pandaized(estimator, X, y, **kwargs):
return f(estimator, np.array(X), safeFactorize(y), **kwargs)
return pandaized
@pandaize
def cross_val_score_pd(estimator, X, y, **kwargs):
return model_selection.cross_val_score(
estimator, X, y, **kwargs)
knnClass = neighbors.KNeighborsClassifier(n_neighbors=3)
cvSchedules = {k : ShuffleSplit(n_splits = 5,
test_size = 0.2,
random_state = 123)
for k in xnorms}
knnCvAccs = {k : np.mean(cross_val_score_pd(estimator = knnClass,
X = xnorms[k],
y = ys[k],
cv = cvSchedules[k].split(xnorms[k])))
for k in xnorms}
## -----------------------------------------------------------------
## try with univariate filter feature selection
## -----------------------------------------------------------------
fsKnnFitter = pipeline.Pipeline([
('featsel', feature_selection.SelectKBest(
feature_selection.f_regression, k=10)),
('classifier', neighbors.KNeighborsClassifier(
n_neighbors=3))
])
fsKnnCvAccs = {k : np.mean(cross_val_score_pd(estimator = fsKnnFitter,
X = xnorms[k],
y = ys[k],
cv = cvSchedules[k].split(xnorms[k])))
for k in xnorms}
## -----------------------------------------------------------------
## vary number of features used
## -----------------------------------------------------------------
nFeatures = [1, 2, 5, 10, 20, 50, 100, 200, 500,
1000, 2000, 5000, 10000]
def fitKnnWithNFeat(n, setname, cv=None):
if cv is None:
cv = cvSchedules[setname]
if n > xnorms[setname].shape[1]:
return None
fsKnnFitter = pipeline.Pipeline([
('featsel', feature_selection.SelectKBest(
feature_selection.f_regression, k=n)),
('classifier', neighbors.KNeighborsClassifier(
n_neighbors=3))
])
return np.mean(cross_val_score_pd(estimator = fsKnnFitter,
X = xnorms[setname],
y = ys[setname],
cv = cv.split(xnorms[setname])))
accsByNFeats = OrderedDict([(s, OrderedDict([(n, fitKnnWithNFeat(n, s))
for n in nFeatures]))
for s in xnorms])
plotData = pd.concat([DataFrame({"set" : s,
"p" : p,
"acc" : accsByNFeats[s][p]},
index = [s + "_" + str(p)])
for s in accsByNFeats
for p in accsByNFeats[s]],
axis = 0)
plotData['acc'] = plotData['acc'].astype(float)
plt.close()
ggo = gg.ggplot(plotData, gg.aes(x='p', y='acc', color='set'))
ggo += gg.geom_line()
ggo += gg.scale_x_log10()
ggo += gg.theme_bw()
print(ggo)
# plotData.to_csv("KnnRealAccuracyByNFeat.tsv",
# sep = "\t",
# index = False,
# header = True)
## -----------------------------------------------------------------
## use PCA feature extraction
## -----------------------------------------------------------------
feKnnFitter = pipeline.Pipeline([
('featextr', pcaextractor.PcaExtractor(k=3)),
('classifier', neighbors.KNeighborsClassifier(
n_neighbors=3))
])
xmod = feKnnFitter.fit(np.array(xnorms['patel']), np.array(ys['patel']))
xcv = cross_val_score_pd(feKnnFitter, xnorms['patel'], ys['patel'],
cv=cvSchedules['patel'].split(xnorms['patel']))
feKnnCvAccs = {k : np.mean(cross_val_score_pd(estimator = feKnnFitter,
X = xnorms[k],
y = ys[k],
cv = cvSchedules[k].split(xnorms[k])))
for k in xnorms}
## -----------------------------------------------------------------
## test with varying number of principal components
## -----------------------------------------------------------------
npcs = [1, 2, 5, 10, 20, 50, 100, 200]
def fitKnnWithNPcs(n, setname, cv=None):
if cv is None:
cv = cvSchedules[setname]
if n > min(xnorms[setname].shape):
return None
feKnnFitter = pipeline.Pipeline([
('featextr', pcaextractor.PcaExtractor(k=n)),
('classifier', neighbors.KNeighborsClassifier(
n_neighbors=3))
])
return np.mean(cross_val_score_pd(estimator = feKnnFitter,
X = xnorms[setname],
y = ys[setname],
cv = cv.split(xnorms[setname])))
accsByNPcs = OrderedDict([(s, OrderedDict([(n, fitKnnWithNPcs(n, s))
for n in npcs]))
for s in xnorms])
plotData = pd.concat([DataFrame({"set" : s,
"p" : p,
"acc" : accsByNPcs[s][p]},
index = [s + "_" + str(p)])
for s in accsByNPcs
for p in accsByNPcs[s]],
axis = 0)
plotData['acc'] = plotData['acc'].astype(float)
plt.close()
ggo = gg.ggplot(plotData, gg.aes(x='p', y='acc', color='set'))
ggo += gg.geom_line()
ggo += gg.scale_x_log10()
ggo += gg.theme_bw()
print(ggo)