/
classif-forest.py
executable file
·158 lines (144 loc) · 5.38 KB
/
classif-forest.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
#!/usr/bin/python
'''Second attempt at disease classification'''
# set global variables in windows to be
# classif2.py -i 0.25 -p 0.05 -n 3 -s -o test "E:\csv"
import numpy as np
import Data
from Query import denest
from sklearn import svm
from sklearn.ensemble import RandomForestClassifier
import argparse,random,csv,os
def sampleReshuffle(dat):
newsamples = random.sample(dat.samples,len(dat.samples))
return newsamples
def crossValIteration(dat,classes,cutoff,prop=0.9,reshuffle=False):
if reshuffle:
dat.samples = sampleReshuffle(dat)
saved_samples = [i for i in dat.samples]
dat.samples = ["{0}_$$_{1}".format(i,v) for i,v in enumerate(dat.samples)]
train,test=dat.splitTraining(prop, classes)
print test.samples
selectedSampleIndicies = [int(i.split("_$$_")[0]) for i in test.samples]
dat.samples = saved_samples
print test.samples
test.samples = [i.split("_$$_")[1] for i in test.samples]
train.samples = [i.split("_$$_")[1] for i in train.samples]
print "Training set has {0} samples from classes: {1}".format(len(train.samples),",".join(set(train.samples)))
print "Test set has {0} samples from classes: {1}".format(len(test.samples),",".join(set(test.samples)))
print "Selecting data..."
# select features for each disease
print "Number of selections made for each class:"
print "Setting up SVM..."
Xtrain = train.values.transpose()
Ytrain = train.samples
clf=RandomForestClassifier(n_estimators=1000)
clf.fit(Xtrain,Ytrain)
Xtest = test.values.transpose()
Ytest = test.samples
print "Predicting R-forest..."
#classification results versus actual
acc = zip(Ytest,clf.predict(Xtest)) # (actual,predicted)... for each sample
print acc # this is the elemental form of the "result" lists processed below
print sum([i[0] == i[1] for i in acc])*1.0/len(acc)
return acc
def dictize(result): # result is a bunch of acc's for iterations of the script above
out ={}
for i in result:
try:
out[i[0]]
except KeyError:
out[i[0]]=[]
out[i[0]].append(i[1])
return out
def sensitivity(resDict):
if not isinstance(resDict,dict):
resDict = dictize(resDict)
out = {}
for i in resDict:
sens = len([j for j in resDict[i] if j==i])*1.0/len(resDict[i])
out[i] = sens
return out
def specificity(resDict):
if not isinstance(resDict,dict):
resDict = dictize(resDict)
out = {}
outDenom = {}
for i in resDict:
out[i]=0
outDenom[i] = 0
for j in resDict:
if not i == j:
out[i]+=len([k for k in resDict[j] if not k == i])
outDenom[i]+=len(resDict[j])
for i in out:
out[i]=out[i]*1./outDenom[i]
return out
def compress(resDicts):
'''For parsing sensitivity and specifictiy results'''
mkLists = lambda x: {i:[x[i]]for i in x}
resDicts = map(mkLists,resDicts)
comp = lambda x,y: {i:x[i]+y[i] for i in x}
return reduce(comp,resDicts)
def makeTable(x):
"Turn a list of dicts with same keys into a matrix"
columns = list(set([j for i in x for j in i.keys()]))
out = []
for row in x:
row_build = []
for col in columns:
row_build.append(row[col])
out.append(row_build)
out = [columns] + out
return out
def doStats(res,format=True):
specif = map(specificity,res)
sensid = map(sensitivity,res)
spec = compress(specif)
sens = compress(sensid)
print sensid,specif
meansd = lambda x: (np.mean(x),np.std(x))
specsd = [meansd(spec[i]) for i in spec]
senssd = [meansd(sens[i]) for i in sens]
out = zip(spec.keys(),specsd,senssd)
if not format:
return out
else:
txtOut = "Disease\tSpecificity(Mean)\tSpecificity(SD)\tSensitivity(Mean)\tSensitivity(SD)\n"
txtOut+="\n".join(["{0}\t{1:.3}\t{2:.3}\t{3:.3}\t{4:.3}".format(i[0],i[1][0],i[1][1],i[2][0],i[2][1]) for i in out])
return txtOut,makeTable(specif),makeTable(sensid)
#expects input of peptides for each class, calculates overlap between them
def pepsInCommon(selPeps):
f = lambda x,y: len([i for i in x if i in y])
return [[f(j) for j in selPeps] for i in selPeps]
def pvalTable(pvals,classes):
l = len(pvals[0])
pv2 = [[j[i] for j in pvals] for i in range(l)] # by class (transpose)t
print pv2
out = {cls:dat for cls,dat in zip(classes,pv2)}
return out
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("infile")
parser.add_argument("-i", "--proportion",type=float,required=True)
parser.add_argument("-p", "--p_value",type=float,required=True)
parser.add_argument("-n", "--num_iterations",type=int,required=True)
parser.add_argument("-o", "--output",required=True)# output directory
parser.add_argument("-s", "--shuffle",action="store_true")
args = parser.parse_args()
print "Importing Data..."
dat=Data.ChipData.fromFileName(args.infile) # read and normalize raw data?
print "Setting classes..."
classes = set(dat.samples)
classes = [i for i in classes]
print "Splitting data..."
acc = [crossValIteration(dat,classes,args.p_value,args.proportion,reshuffle=args.shuffle) for i in range(args.num_iterations)]
res,spec,sens = doStats(acc)
os.mkdir(args.output)
os.chdir(args.output)
import csv
with open("sensitivity","w") as f:
wtr = csv.writer(f)
wtr.writerows(sens)
with open("specificity","w") as f:
wtr = csv.writer(f)
wtr.writerows(spec)