测试新数据预测

Question

我已经使用 Yelp-Data-Challenge 训练了一个模型，数据得到了一个 pickle 文件 399850by50reviews_words_index.pkl，但我一直坚持如何使用这个 pickle 文件来测试 keras 中的新数据

这是我用于训练数据并保存到模型创建中的代码

如何将此模型用于测试数据

我在这里使用 Keras 1.0.0 和 Theano

'''
train cnn mode for sentiment classification on yelp data set
author: hao peng
'''
import pickle
import pandas as pd
import numpy as np
from sklearn.cross_validation import train_test_split
from Word2VecUtility import Word2VecUtility
from keras.preprocessing import sequence
from keras.models import Sequential
from keras.layers.core import Dense, Dropout, Activation, Flatten
from keras.layers.embeddings import Embedding
from keras.layers.convolutional import Convolution1D, MaxPooling1D


def get_volcabulary_and_list_words(data):
    reviews_words = []
    volcabulary = []
    for review in data["text"]:
        review_words = Word2VecUtility.review_to_wordlist(
            review, remove_stopwords=True)
        reviews_words.append(review_words)
        for word in review_words:
            volcabulary.append(word)
    volcabulary = set(volcabulary)
    return volcabulary, reviews_words

def get_reviews_word_index(reviews_words, volcabulary, max_words, max_length):
    word2index = {word: i for i, word in enumerate(volcabulary)}
    # use w in volcabulary to limit index within max_words
    reviews_words_index = [[start] + [(word2index[w] + index_from) for w in review] for review in reviews_words]
    # in word2vec embedding, use (i < max_words + index_from) because we need the exact index for each word, in order to map it to its vector. And then its max_words is 5003 instead of 5000.
    reviews_words_index = [[i if (i < max_words) else oov for i in index] for index in reviews_words_index]
    # padding with 0, each review has max_length now.
    reviews_words_index = sequence.pad_sequences(reviews_words_index, maxlen=max_length, padding='post', truncating='post')
    return reviews_words_index

def vectorize_labels(labels, nums):
    labels = np.asarray(labels, dtype='int32')
    length = len(labels)
    Y = np.zeros((length, nums))
    for i in range(length):
        Y[i, (labels[i]-1)] = 1.
    return Y
# data processing para
max_words = 5000
max_length = 50

# model training parameters
batch_size = 32
embedding_dims = 100
nb_filter = 250
filter_length = 3
hidden_dims = 250
nb_epoch = 2

# index trick parameters
index_from = 3
start = 1
# padding = 0
oov = 2

data = pd.read_csv(
    'review_sub_399850.tsv', header=0, delimiter="\t", quoting=3, encoding='utf-8')
print('get volcabulary...')
volcabulary, reviews_words = get_volcabulary_and_list_words(data)
print('get reviews_words_index...')
reviews_words_index = get_reviews_word_index(reviews_words, volcabulary, max_words, max_length)

print reviews_words_index[:20, :12]
print reviews_words_index.shape

labels = data["stars"]

pickle.dump((reviews_words_index, labels), open("399850by50reviews_words_index.pkl", 'wb'))

(reviews_words_index, labels) = pickle.load(open("399850by50reviews_words_index.pkl", 'rb'))

index = np.arange(reviews_words_index.shape[0])
train_index, valid_index = train_test_split(
    index, train_size=0.8, random_state=100)

labels = vectorize_labels(labels, 5)
train_data = reviews_words_index[train_index]
valid_data = reviews_words_index[valid_index]
train_labels = labels[train_index]
valid_labels = labels[valid_index]
print train_data.shape
print valid_data.shape
print train_labels[:10]

del(labels, train_index, valid_index)

print "start training model..."

model = Sequential()

# we start off with an efficient embedding layer which maps
# our vocab indices into embedding_dims dimensions
model.add(Embedding(max_words + index_from, embedding_dims, \
                    input_length=max_length))
model.add(Dropout(0.25))

# we add a Convolution1D, which will learn nb_filter
# word group filters of size filter_length:

# filter_length is like filter size, subsample_length is like step in 2D CNN.
model.add(Convolution1D(nb_filter=nb_filter,
                        filter_length=filter_length,
                        border_mode='valid',
                        activation='relu',
                        subsample_length=1))
# we use standard max pooling (halving the output of the previous layer):
model.add(MaxPooling1D(pool_length=2))

# We flatten the output of the conv layer,
# so that we can add a vanilla dense layer:
model.add(Flatten())

# We add a vanilla hidden layer:
model.add(Dense(hidden_dims))
model.add(Dropout(0.25))
model.add(Activation('relu'))

# We project onto 5 unit output layer, and activate it with softmax:
model.add(Dense(5))
model.add(Activation('softmax'))

model.compile(loss='categorical_crossentropy',
              optimizer='sgd',
              class_mode='categorical')
model.fit(train_data, train_labels, batch_size=batch_size,
          nb_epoch=nb_epoch, show_accuracy=True,
          validation_data=(valid_data, valid_labels))

Answer 1

您用于测试的输入数据的形状必须与train_data 和valid_data 完全相同，除了第一个维度是批量大小。

因此，您必须使用要测试的输入数据创建一个 numpy 数组，并确保该数组的结构与train_data 完全相同，即yourTestArray.shape[1:] 与train_data.shape[1:] 完全相同，也等于valid_data.shape[1:]。

拥有该数组后，您应该使用results = model.predict(yourTestArray)。