from __future__ import division from sklearn import ensemble import sklearn import windowfile import numpy as np import matplotlib.pyplot as plt import sys import pickle print('load') A = windowfile.readwins(open(sys.argv[1])) B = windowfile.readwins(open(sys.argv[2])) tamA = A.shape[0] tamB = B.shape[0] trainingSize = 10000 singleSet = np.zeros((trainingSize, 256)) overlapSet = np.zeros((trainingSize, 256)) print('generate set') for i in xrange(trainingSize): # single sA = int(np.random.rand() * tamA) sB = int(np.random.rand() * tamB) if np.random.rand() > 0.5: # from A singleSet[i, :] = A[sA] else: singleSet[i, :] = B[sA]
from __future__ import division import sys import svm import windowfile import pickle import numpy as np import matplotlib.pyplot as plt import scipy.signal as sg print('load model') model = svm.libsvm.svm_load_model('/ssd/15o04000_15o04001_h.svmmodel') print('load A') A = windowfile.readwins(open('/ssd/15o03000_h.features')) print('load sig') sig = windowfile.readwinsEx(open('/ssd/15o03000_h.spikes')) print('load randomForest') clf = pickle.load(open("RandomForestOverlapModel.pickle", 'rb')) state = 'single' probs = (svm.c_double * 2)(0, 0) pA = 1. pB = 1. maxamp = 0. sigs_now = [] #H = np.zeros(256) print('-') counter = 0 tam = A.shape[0] for i in xrange(tam): if i % 10000 == 0:
from __future__ import division import sys import svm import windowfile import pickle import numpy as np import matplotlib.pyplot as plt import scipy.signal as sg print('load model') model = svm.libsvm.svm_load_model('/ssd/15o04000_15o04001_h.svmmodel') print('load A') A = windowfile.readwins(open('/ssd/15o03000_h.features')) print('load sig') sig = windowfile.readwinsEx(open('/ssd/15o03000_h.spikes')) print('load randomForest') clf = pickle.load( open("RandomForestOverlapModel.pickle", 'rb')) state = 'single' probs = (svm.c_double*2)(0,0) pA = 1. pB = 1. maxamp = 0. sigs_now = [] #H = np.zeros(256) print('-') counter = 0 tam = A.shape[0] for i in xrange(tam):
import sys import numpy as np import matplotlib.pyplot as plt import windowfile # | awk '{print $4" "$2 }' | sort A = windowfile.readwins(open(sys.argv[1])) B = windowfile.readwins(open(sys.argv[2])) assert (A.shape[1] == B.shape[1]) for i in xrange(A.shape[1]): plt.clf() pdf, bins, patches = plt.hist((A[:, i], B[:, i]), bins=50, normed=True) minpdf = np.min(np.vstack(pdf), axis=0) overlap = np.sum(minpdf * np.diff(bins)) print('feature %4d overlap: %.10f' % (i, overlap)) sys.stdout.flush() plt.savefig('out/fea%04d.png' % i)