Run Random Forest on all data: training and the best test result so far
In [1]:
%matplotlib inline
from matplotlib import pylab as pl
import cPickle as pickle
import pandas as pd
import numpy as np
import os
import random
In [2]:
import sys
sys.path.append('..')
uncommoent the relevant pipeline in ../seizure_detection.py and run
cd ..
./doall data
or
./doall td
./doall tt
In [3]:
FEATURES = 'gen-8_medianwindow-bands2-usf-w60-b0.2-b4-b8-b12-b30-b70-0.1-0.5-0.9'
In [4]:
from common.data import CachedDataLoader
cached_data_loader = CachedDataLoader('../data-cache')
In [5]:
def read_data(target, data_type, features):
fname = 'data_%s_%s_%s'%(data_type,target,features)
print fname
return cached_data_loader.load(fname,None)
In [6]:
best = pd.read_csv('../submissions/141029-predict.10.csv', index_col='clip', squeeze=True)
In [7]:
best.hist(bins=50)
Out[7]:
In [8]:
from sklearn.ensemble import RandomForestClassifier
from sklearn.cross_validation import StratifiedKFold
from sklearn.metrics import roc_auc_score
from sklearn.linear_model import LogisticRegression as LR
clf = RandomForestClassifier(n_estimators=3000, min_samples_split=1, bootstrap=False,max_depth=10,
n_jobs=-1)#, max_features=15
In [9]:
fpout = open('../submissions/141101-predict.3.csv','w')
print >>fpout,'clip,preictal'
In [10]:
def prb2logit(x):
return np.log(x/(1.-x))
def logit2prb(x):
return 1./(1+np.exp(-x))
SMOOTH = 0.3
for target in ['Dog_1', 'Dog_2', 'Dog_3', 'Dog_4', 'Dog_5', 'Patient_1', 'Patient_2']:
pdata0 = read_data(target, 'preictal', FEATURES) # positive examples
ndata0 = read_data(target, 'interictal', FEATURES) # negative examples
X0 = np.concatenate((pdata0.X, ndata0.X))
y0 = np.zeros(X0.shape[0])
y0[:pdata0.X.shape[0]] = 1
# predict
tdata = read_data(target, 'test', FEATURES) # test examples
Xt = tdata.X
Nt = Xt.shape[0]
yt = np.array([best['%s_test_segment_%04d.mat' % (target, i+1)] for i in range(Nt)])
yt = prb2logit(np.clip(yt,0.01,0.99))
yt = yt*(1.-SMOOTH) + SMOOTH*prb2logit(y0.mean())
yt = logit2prb(yt)
X = np.concatenate((X0,Xt))
y_proba = []
for j in range(10):
y1 = np.array([np.random.random(Nt) < yt],dtype=int).ravel()
y = np.concatenate((y0,y1))
clf.fit(X,y)
y_proba.append(clf.predict_proba(Xt)[:,1])
y_proba = np.array(y_proba).mean(axis=0)
# write results
for i,p in enumerate(y_proba):
print >>fpout,'%s_test_segment_%04d.mat,%.15f' % (target, i+1, p)
In [11]:
fpout.close()