【CNN文本分类调用数据集】

偏偏宠爱

已于 2023-12-20 18:02:35 修改

阅读量89

点赞数

分类专栏：文本分类文章标签： tensorflow 深度学习 python

于 2023-02-26 22:40:19 首次发布

本文链接：https://blog.csdn.net/Eternity68/article/details/129232989

版权

文本分类专栏收录该内容

2 篇文章 0 订阅

订阅专栏

CNN文本分类

import warnings
warnings.filterwarnings("ignore")
import numpy as np
#import matplotlib.pylot as plt
from tensorflow import keras
from tensorflow.keras import layers
from tensorflow.keras.preprocessing.sequence import pad_sequences

num_features=3000
sequence_length=300
embedding_dimension=100
(x_train,y_train),(x_test,y_test)=keras.datasets.imdb.load_data(num_words=num_features)
print(x_train.shape)
print(x_test.shape)
print(y_train.shape)
print(y_test.shape)

run

(25000,)
(25000,)
(25000,)
(25000,)

x_train=pad_sequences(x_train,maxlen=sequences_length)
x_test=pad_sequences(x_test,maxlen=sequences_length)
print(x_train.shape)
print(x_test.shape)
print(y_train.shape)
print(y_test.shape)

run

(25000, 300)
(25000, 300)
(25000,)
(25000,)

filter_size=[3,4,5]
def convolution():
    inn=layers.Input(shape=(sequences_length,embedding_dimension,1))
    cnns=[]
    for size in filter_size:
        conv=layers.Conv2D(filters=64,kernel_size=(size,embedding_dimension),strides=1,padding='valid',activation='relu')(inn)
        pool=layers.MaxPool2D(pool_size=(sequence_length-size+1,1),padding='valid')(conv)
        cnns.append(pool)
    outt=layers.concatenate(cnns)
    model=keras.Model(inputs=inn,outputs=outt)
    return model

def cnn_mulfilter():
    model=keras.Sequential([layers.Embedding(input_dim=num_features,output_dim=embedding_dimension,input_length=sequence_length),
                           layers.Reshape((sequence_length,embedding_dimension,1)),
                           convolution(),
                           layers.Flatten(),
                           layers.Dense(10,activation='relu'),
                           layers.Dropout(0.2),
                           layers.Dense(1,activation='sigmoid')])
    model.compile(optimizer=keras.optimizers.Adam(),loss=keras.losses.BinaryCrossentropy(),metrics=['accuracy'])
    return model

model=cnn_mulfilter()
model.summary()

run

Model: “sequential_1”

Layer (type) Output Shape Param #

embedding_4 (Embedding) (None, 300, 100) 300000

reshape_4 (Reshape) (None, 300, 100, 1) 0

model_1 (Functional) (None, 1, 1, 192) 76992

flatten_1 (Flatten) (None, 192) 0

dense_2 (Dense) (None, 10) 1930

dropout_1 (Dropout) (None, 10) 0

dense_3 (Dense) (None, 1) 11

Total params: 378,933
Trainable params: 378,933
Non-trainable params: 0

history=model.fit(x_train,y_train,batch_size=64,epochs=5,validation_split=0.1)

run

Epoch 1/5
352/352 [] - 34s 44ms/step - loss: 0.4525 - accuracy: 0.7779 - val_loss: 0.2980 - val_accuracy: 0.8820
Epoch 2/5
352/352 [] - 9s 26ms/step - loss: 0.2624 - accuracy: 0.8979 - val_loss: 0.2695 - val_accuracy: 0.8928
Epoch 3/5
352/352 [] - 9s 26ms/step - loss: 0.1841 - accuracy: 0.9348 - val_loss: 0.2848 - val_accuracy: 0.8912
Epoch 4/5
352/352 [] - 9s 26ms/step - loss: 0.1242 - accuracy: 0.9611 - val_loss: 0.3109 - val_accuracy: 0.8872
Epoch 5/5
352/352 [==============================] - 9s 26ms/step - loss: 0.0760 - accuracy: 0.9798 - val_loss: 0.3482 - val_accuracy: 0.8876