this code runs LR and NN at the same time """ import numpy as np import IOutils # from sklearn.naive_bayes import GaussianNB from sklearn.linear_model import LogisticRegression from sklearn import svm import random import sys from sklearn.metrics import confusion_matrix, precision_score, recall_score, f1_score, classification_report from neuralnetworks.templates import BasicNN2 nn = BasicNN2(max_epochs=100, hidden=[200,50,30], input_shape=(None, 42), output_num_units=2) ds = IOutils.data_streamer2(keeplist=[ (i,0) for i in xrange(1,12) ]) vt = IOutils.VectorTransformer() # NaivB = GaussianNB() lr = LogisticRegression(class_weight="auto") #svc = svm.SVC(kernel='rbf',C=10,class_weight="auto") #======= # svc = svm.SVC(kernel='rbf',C=10,class_weight="auto") for X_next, Y_next in ds: X_next = X_next.astype(np.float)/X_next.max() X_next[np.isnan(X_next)]=0 zipped = zip(X_next,Y_next) random.shuffle(zipped) X,Y = zip(*zipped)
# In[1]:
import numpy as np
import matplotlib.pylab as plt
from collections import Counter
import IOutils
import random
from neuralnetworks.templates import BasicNN2
from sklearn.metrics import confusion_matrix, classification_report
from itertools import product
# In[2]:
training_ds = IOutils.data_streamer2(keeplist=[ (i, 0) for i in range(1,12) ]) # obtain data for first 11 patients
nn = BasicNN2(input_shape=(None,42), output_num_units=11, max_epochs=100, hidden=[200,120,30])
vt = IOutils.VectorTransformer()
# In[3]:
n_repeat_sampling = 1
dataset_count = 0
for X,Y in training_ds:
dataset_count += 1
# transform the Ys
Y = vt.transform(Y)
# print('total size before sampling:', len(Y))
import matplotlib.pylab as plt
from collections import Counter
import sys
if not len(sys.argv)>1:
print """arguments:
SUBJECT_ID: Id of the subject you want to train
"""
raise Exception('NEED MORE ARGUMENTS')
# NUM_ZERO_METRIC = sys.argv[3]
# obtain the first 7 datas for the 2nd subject
subject_id = int(sys.argv[1])
training_ds = IOutils.data_streamer2(keeplist=[ (subject_id, i) for i in range(1,7) ])
#nn = BasicNN(input_shape=(None,42), output_num_units=12, max_epochs=int(sys.argv[5]), hidden_num_units=int(sys.argv[4]))
vt = IOutils.VectorTransformer()
linear = LogisticRegression(class_weight = 'auto')
# n_repeat_sampling = int(sys.argv[2])
dataset_count = 0
for X,Y in training_ds:
dataset_count += 1
# transform the Ys
Y = vt.transform(Y)
# print('total size before sampling:', len(Y))
X = X.astype(np.float)
# normalization for regression
X[np.isnan(X)] = 0
import IOutils # from sklearn.naive_bayes import GaussianNB from sklearn.linear_model import LogisticRegression from sklearn.ensemble import RandomForestClassifier # from sklearn import svm import random import sys from sklearn.metrics import confusion_matrix from sklearn.metrics import precision_score from sklearn.metrics import recall_score from sklearn.metrics import f1_score ds = IOutils.data_streamer2() vt = IOutils.VectorTransformer() X_valid, Y_valid = ds.next() # use as follows: Y_valid = vt.transform(Y_valid) Y_valid [Y_valid != 0] = 1 # Y_valid [Y_valid == 0] = 0 # NaivB = GaussianNB() # linear = LogisticRegression(class_weight = 'auto') #accounts for class imbalance # support = svm.SVC(kernel='rbf',C=10)
import numpy as np
import os
import pandas as pd
# Grab train data
data = []
label = []
n_sub = 1
n_series = 8
train_streamer = io.data_streamer2(mode='train')
for k in range(n_sub):
sub_data = []
sub_label = []
for series in range(n_series):
d, e = train_streamer.next()
sub_data.append(d)
sub_label.append(e)
data.append(sub_data)
label.append(sub_label)
np.save('eeg_train.npy', [data, label])
del data, label
# coding: utf-8
# In[1]:
import numpy as np
import matplotlib.pylab as plt
from collections import Counter
import IOutils
import random
from neuralnetworks.templates import BasicNN2
from sklearn.metrics import confusion_matrix, classification_report
# In[ ]:
training_ds = IOutils.data_streamer2(keeplist=[ (1, i) for i in range(1,7) ]) # obtain the first 7 datas for the 1st subject
nn = BasicNN2(input_shape=(None,42), output_num_units=11, max_epochs=100, hidden=[200,120,30])
vt = IOutils.VectorTransformer()
# In[ ]:
n_repeat_sampling = 1
dataset_count = 0
for X,Y in training_ds:
dataset_count += 1
# transform the Ys
Y = vt.transform(Y)
# print('total size before sampling:', len(Y))
X = X.astype(np.float)
# normalization for regression
