我的excel文件中有四万六千条标注好了positive、neutral、negative三种情感的数据，给我卷积神经网络文本情感三分类的代码，除了分训练集和测试集输出报告，还要输出他们的准确率

好的，以下是一个基于Keras的卷积神经网络文本情感三分类的代码示例：

import pandas as pd
import numpy as np
from keras.layers import Dense, Input, Conv1D, MaxPooling1D, Flatten, Dropout, Embedding
from keras.models import Model
from keras.utils import to_categorical
from sklearn.metrics import classification_report
from sklearn.model_selection import train_test_split
from keras.preprocessing.text import Tokenizer
from keras.preprocessing.sequence import pad_sequences


# 读取数据
df = pd.read_excel('data.xlsx')
texts = df['text'].values
labels = df['label'].values

# 将标签转换为独热编码
labels = to_categorical(labels)

# 对文本进行分词和编码
MAX_NUM_WORDS = 20000
MAX_SEQUENCE_LENGTH = 100
tokenizer = Tokenizer(num_words=MAX_NUM_WORDS)
tokenizer.fit_on_texts(texts)
sequences = tokenizer.texts_to_sequences(texts)
word_index = tokenizer.word_index
data = pad_sequences(sequences, maxlen=MAX_SEQUENCE_LENGTH)

# 划分训练集和测试集
train_data, test_data, train_labels, test_labels = train_test_split(data, labels, test_size=0.2, random_state=42)

# 定义模型结构
embedding_dim = 100
inputs = Input(shape=(MAX_SEQUENCE_LENGTH,), dtype='int32')
embedding_layer = Embedding(MAX_NUM_WORDS, embedding_dim, input_length=MAX_SEQUENCE_LENGTH)(inputs)
conv1 = Conv1D(filters=128, kernel_size=3, activation='relu')(embedding_layer)
pool1 = MaxPooling1D(pool_size=2)(conv1)
conv2 = Conv1D(filters=64, kernel_size=3, activation='relu')(pool1)
pool2 = MaxPooling1D(pool_size=2)(conv2)
flatten = Flatten()(pool2)
dense1 = Dense(64, activation='relu')(flatten)
dropout1 = Dropout(0.5)(dense1)
dense2 = Dense(32, activation='relu')(dropout1)
dropout2 = Dropout(0.5)(dense2)
outputs = Dense(3, activation='softmax')(dropout2)

model = Model(inputs=inputs, outputs=outputs)

# 编译模型
model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

# 训练模型
model.fit(train_data, train_labels, batch_size=64, epochs=10, validation_data=(test_data, test_labels))

# 在测试集上评估模型
y_pred = np.argmax(model.predict(test_data), axis=1)
y_true = np.argmax(test_labels, axis=1)
print(classification_report(y_true, y_pred))

# 输出准确率
accuracy = np.sum(y_pred == y_true) / len(y_true)
print("Accuracy:", accuracy)

需要注意的是，这里使用了Keras自带的Tokenizer对文本进行了分词和编码，使用了Embedding层将文本转换为向量表示。在训练模型时，我们使用了Adam优化器和交叉熵损失函数，评估指标为准确率。在测试集上评估模型时，我们使用sklearn库中的classification_report函数输出分类报告，同时计算并输出准确率。