ameasure · March 1, 2016 13:36 · ameasure · Mar 1, 2016
diff --git a/lstm.py b/lstm.py
 import datetime

 import pandas as pd
 import numpy as np
 np.random.seed(42)

 from sklearn.preprocessing import LabelEncoder

 import theano
 import keras
 from keras.models import Sequential, Graph
 from keras.layers.convolutional import Convolution1D, MaxPooling1D, Convolution2D, MaxPooling2D
 from keras.layers.core import Dense, Dropout, Activation, Flatten, Reshape, Merge
 from keras.layers.recurrent import LSTM
 from keras.layers.normalization import BatchNormalization
 from keras.layers.embeddings import Embedding
 from keras.optimizers import SGD
 from keras.utils import np_utils
 from keras.preprocessing.text import Tokenizer
 from keras.preprocessing.sequence import pad_sequences
 from keras.regularizers import l2

 from embed_preprocessing import get_initial_embeddings
 from msha_extractor import get_data

 max_len = 100
 embedding_size = 300
 batch_size = 128

 code_type = 'ACTIVITY_CD'
 raw_train, raw_valid, raw_test = get_data(n_train=50000, n_valid=0, n_test=10000)
 raw_train_labels = raw_train[code_type]
 raw_valid_labels = raw_valid[code_type]
 raw_test_labels = raw_test[code_type]

 labeler = LabelEncoder()
 labels = set(raw_train[code_type].tolist() + raw_test[code_type].tolist())
 labeler.fit(list(labels))
 nb_classes = len(set(labels))
 print('nb_classes = %s' % nb_classes)

 y_train = labeler.transform(raw_train_labels)
 Y_train = np_utils.to_categorical(y_train, nb_classes)

 y_valid = labeler.transform(raw_valid_labels)
 Y_valid = np_utils.to_categorical(y_valid, nb_classes)

 y_test = labeler.transform(raw_test_labels)
 Y_test = np_utils.to_categorical(y_test, nb_classes)

 print 'Tokenizing X_train'
 tokenizer = Tokenizer()
 tokenizer.fit_on_texts(raw_train['NARRATIVE'])
 tokenizer.word_index['BLANK_EMBEDDING'] = 0
 X_train = tokenizer.texts_to_sequences(raw_train['NARRATIVE'])
 X_valid = tokenizer.texts_to_sequences(raw_valid['NARRATIVE'])
 X_test = tokenizer.texts_to_sequences(raw_test['NARRATIVE'])
 X_train = pad_sequences(X_train, maxlen=max_len)
 X_valid = pad_sequences(X_valid, maxlen=max_len)
 X_test = pad_sequences(X_test, maxlen=max_len)
 print('X_train shape:', X_train.shape)
 print('X_valid shape:', X_valid.shape)
 print('X_test shape:', X_test.shape)

 initial_embeddings = get_initial_embeddings(tokenizer.word_index)
 vocab_size = len(tokenizer.word_index)
 filter_length = 3
 nb_filter = 300

 model = Sequential()
 model.add(Embedding(vocab_size, embedding_size, input_length=max_len, weights=[initial_embeddings]))
 #model.add(Convolution1D(nb_filter=nb_filter,
 #                        filter_length=filter_length,
 #                        border_mode='valid',
 #                        activation='relu',
 #                        subsample_length=1))
 model.add(LSTM(150, dropout_W=0.5, dropout_U=0.5, return_sequences=True))
 model.add(MaxPooling1D(pool_length=3))
 model.add(Flatten())
 model.add(Dropout(0.5))
 model.add(Dense(output_dim=nb_classes))
 model.add(Activation('sigmoid'))

 model.compile(loss='categorical_crossentropy',
              optimizer='adam')
              
 model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=5, 
          verbose=True, show_accuracy=True, validation_split=.05)

 score = model.evaluate(X_test, Y_test, batch_size=batch_size, show_accuracy=True)
 print('Test score:', score[0])
 print('Test accuracy:', score[1])
	import datetime

	import pandas as pd
	import numpy as np
	np.random.seed(42)

	from sklearn.preprocessing import LabelEncoder

	import theano
	import keras
	from keras.models import Sequential, Graph
	from keras.layers.convolutional import Convolution1D, MaxPooling1D, Convolution2D, MaxPooling2D
	from keras.layers.core import Dense, Dropout, Activation, Flatten, Reshape, Merge
	from keras.layers.recurrent import LSTM
	from keras.layers.normalization import BatchNormalization
	from keras.layers.embeddings import Embedding
	from keras.optimizers import SGD
	from keras.utils import np_utils
	from keras.preprocessing.text import Tokenizer
	from keras.preprocessing.sequence import pad_sequences
	from keras.regularizers import l2

	from embed_preprocessing import get_initial_embeddings
	from msha_extractor import get_data

	max_len = 100
	embedding_size = 300
	batch_size = 128

	code_type = 'ACTIVITY_CD'
	raw_train, raw_valid, raw_test = get_data(n_train=50000, n_valid=0, n_test=10000)
	raw_train_labels = raw_train[code_type]
	raw_valid_labels = raw_valid[code_type]
	raw_test_labels = raw_test[code_type]

	labeler = LabelEncoder()
	labels = set(raw_train[code_type].tolist() + raw_test[code_type].tolist())
	labeler.fit(list(labels))
	nb_classes = len(set(labels))
	print('nb_classes = %s' % nb_classes)

	y_train = labeler.transform(raw_train_labels)
	Y_train = np_utils.to_categorical(y_train, nb_classes)

	y_valid = labeler.transform(raw_valid_labels)
	Y_valid = np_utils.to_categorical(y_valid, nb_classes)

	y_test = labeler.transform(raw_test_labels)
	Y_test = np_utils.to_categorical(y_test, nb_classes)

	print 'Tokenizing X_train'
	tokenizer = Tokenizer()
	tokenizer.fit_on_texts(raw_train['NARRATIVE'])
	tokenizer.word_index['BLANK_EMBEDDING'] = 0
	X_train = tokenizer.texts_to_sequences(raw_train['NARRATIVE'])
	X_valid = tokenizer.texts_to_sequences(raw_valid['NARRATIVE'])
	X_test = tokenizer.texts_to_sequences(raw_test['NARRATIVE'])
	X_train = pad_sequences(X_train, maxlen=max_len)
	X_valid = pad_sequences(X_valid, maxlen=max_len)
	X_test = pad_sequences(X_test, maxlen=max_len)
	print('X_train shape:', X_train.shape)
	print('X_valid shape:', X_valid.shape)
	print('X_test shape:', X_test.shape)

	initial_embeddings = get_initial_embeddings(tokenizer.word_index)
	vocab_size = len(tokenizer.word_index)
	filter_length = 3
	nb_filter = 300

	model = Sequential()
	model.add(Embedding(vocab_size, embedding_size, input_length=max_len, weights=[initial_embeddings]))
	#model.add(Convolution1D(nb_filter=nb_filter,
	# filter_length=filter_length,
	# border_mode='valid',
	# activation='relu',
	# subsample_length=1))
	model.add(LSTM(150, dropout_W=0.5, dropout_U=0.5, return_sequences=True))
	model.add(MaxPooling1D(pool_length=3))
	model.add(Flatten())
	model.add(Dropout(0.5))
	model.add(Dense(output_dim=nb_classes))
	model.add(Activation('sigmoid'))

	model.compile(loss='categorical_crossentropy',
	optimizer='adam')

	model.fit(X_train, Y_train, batch_size=batch_size, nb_epoch=5,
	verbose=True, show_accuracy=True, validation_split=.05)

	score = model.evaluate(X_test, Y_test, batch_size=batch_size, show_accuracy=True)
	print('Test score:', score[0])
	print('Test accuracy:', score[1])