文本分類（3）-卷積神經網絡（CNN）實現文本分類

# cnn實現垃圾郵件分類
from keras.models import Sequential
from keras.layers import Dense, Conv1D, GlobalMaxPooling1D, Embedding, Dropout, Activation, MaxPooling1D
from keras.preprocessing.text import Tokenizer
from keras.preprocessing import sequence
from keras import metrics
import keras.preprocessing.text
from sklearn.preprocessing import LabelEncoder
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

data_file = 'spam.csv'

# 讀取數據
df = pd.read_csv(data_file, encoding='latin-1')

# 標籤
labels = df.v1
# 文本
texts = df.v2

# 預處理，將一個句子拆分成單詞構成列表
def text_to_word_sequence(text, filters='!"#$%&()*+,-./:;<=>?@[\\]^_`{|}~\t\n',
                         lower=True, split=" "):
    if lower: 
        text = text.lower()
    if type(text) == unicode:
        translate_table = {ord(c): ord(t) for c, t in zip(filters, split*len(filters))}
    else:
        translate_table = maketrans(filters, split*len(filters))
    text = text.translate(translate_table)
    seq = text.split(split)
    
    return [i for i in seq if i]

keras.preprocessing.text.text_to_word_sequence = text_to_word_sequence

num_max = 1000

le = LabelEncoder()
labels = le.fit_transform(labels)

# 分詞器
tok = Tokenizer(num_words=num_max)  # num_words：處理單詞最大數量
tok.fit_on_texts(texts)

mat_texts = tok.texts_to_matrix(texts, mode='count') # 文本向量化

n_sample = mat_texts.shape[0]

max_len = 1000  # 序列最大長度
cnn_texts_seq = tok.texts_to_sequences(texts)  # 文本轉化爲序列
# 填充序列，短於序列的最大長度用0填充，長於序列最大長度進行截斷
cnn_texts_mat = sequence.pad_sequences(cnn_texts_seq, maxlen=max_len)  

# cnn
def get_cnn_model():
    model = Sequential()
    model.add(Embedding(1000, 20, input_length=max_len))  # 輸入維度（詞彙表大小）；詞向量維度；輸入序列長度
    model.add(Dropout(0.2))
    model.add(Conv1D(32, 5, strides=1, padding='valid', activation='relu'))  # 256
#     model.add(MaxPooling1D())
#     model.add(Conv1D(512, 5, strides=1, padding='valid', activation='relu'))
    model.add(GlobalMaxPooling1D())
    model.add(Dense(64, activation='relu'))  # 64
    model.add(Dropout(0.2))
    model.add(Dense(1, activation='sigmoid'))
    model.summary()
    
    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc'])
    
    return model

# 訓練
def train_model(model, x, y):
    return model.fit(x, y, batch_size=32, epochs=10, verbose=1, validation_split=0.2)

n_validation = int(n_sample*0.8)

# 訓練集
x_train = cnn_texts_mat[:n_validation]
# 標籤
y_train = labels[:n_validation]

# 驗證集
x_val = cnn_texts_mat[n_validation:]
y_val = labels[n_validation:]

m = get_cnn_model()
history = train_model(m, x_train, y_train)

# 評估
def test_model(model, x, y):
    return model.evaluate(x, y)

loss, accuracy = test_model(m, x_val, y_val)
print('loss: ', loss)
print('accuracy: ', accuracy)

# 預測
def predicted(model, x):
    return model.predict_classes(x)

y_pred = predicted(m, x_val)

# print(y_pred[:10])

# 繪圖
plt.figure(figsize=(10, 5))

plt.subplot(121)
plt.plot(history.history['acc'], c='b', label='train')
plt.plot(history.history['val_acc'], c='g', label='validation')
plt.legend()
plt.xlabel('epoch')
plt.ylabel('accuracy')
plt.title('Model accuracy')

plt.subplot(122)
plt.plot(history.history['loss'], c='b', label='train')
plt.plot(history.history['val_loss'], c='g', label='validation')
plt.legend()
plt.xlabel('epoch')
plt.ylabel('loss')
plt.title('Model loss')

plt.show()