利用tensorflow做深度学习情感分析

深度学习作为一项学习数据的多层特征或表征的强大机器学习技术。此项目中，将使用tensorflow深度学习平台，通过相关模型的构建，以及数据的处理，完成微博评论情感分析，已到达类似百度AI情感分析功能，可以通过用户输入的数据进行分析，得到情感为积极性还是消极性。

1. 结构

• html前端界面
• 图像识别模型
• flask后端接口 API

2. 前端HTML设计

前端仿的百度ai的情感分析界面。只是做一个简单的交互页面，使用纯html+js就行了。当然，作为毕设， 你也可以花更多的想法去完善前端界面，毕竟答辩时的第一印象就是前端展示。这里也能做成.exe程序，反正都一样。

3. 模型搭建

3.1 lstm+attention模型

采用这两个来做。说白了就是rnn plus。

class Attention(Layer):
    def __init__(self, step_dim,
                W_regularizer=None, b_regularizer=None,
                W_constraint=None, b_constraint=None,
                bias=True, **kwargs):
        
        self.supports_masking = True
        self.init = initializers.get('glorot_uniform')
        
        self.W_regularizer = regularizers.get(W_regularizer)
        self.b_regularizer = regularizers.get(b_regularizer)

        self.W_constraint = constraints.get(W_constraint)
        self.b_constraint = constraints.get(b_constraint)

        self.bias = bias
        self.step_dim = step_dim

        self.features_dim = 0

        super(Attention, self).__init__(**kwargs)


    def build(self, input_shape):
        assert len(input_shape) == 3              

        self.W = self.add_weight(name='{}_W'.format(self.name), 
                shape=(input_shape[-1],), initializer=self.init,
                regularizer=self.W_regularizer,constraint=self.W_constraint
            )

        self.features_dim = input_shape[-1]

        if self.bias:
            self.b = self.add_weight(name='{}_b'.format(self.name),
                    shape=(input_shape[1],), initializer='zero', 
                    regularizer=self.b_regularizer, constraint=self.b_constraint
                )

        else:
            self.b = None

        self.built = True

    def compute_mask(self, input, input_mask=None):
        return None


    def call(self, x, mask=None):
        features_dim = self.features_dim
        step_dim = self.step_dim

        e = K.reshape(K.dot(K.reshape(x, (-1, features_dim)), K.reshape(self.W, (features_dim, 1))), (-1, step_dim))

        if self.bias:
            e += self.b
        
        e = K.tanh(e)

        a = K.exp(e)

        if mask is not None:
            a *= K.cast(mask, K.floatx())

        a /= K.cast(K.sum(a, axis=1, keepdims=True) + K.epsilon(), K.floatx())
        a = K.expand_dims(a)

        c = K.sum(a * x, axis=1)
        return c

    def compute_output_shape(self, input_shape):
        return input_shape[0], self.features_dim


from tensorflow.keras import Input, Model
from tensorflow.keras.layers import Embedding, Dense, Dropout, Bidirectional, LSTM, GRU

# 双向长短时间记忆模型
# 增加注意力机制
class TextAttBiRNN(object):
    def __init__(self, maxlen, max_features, embedding_dims, class_num = 10, last_cativation = 'sigmoid'):
        self.maxlen = maxlen
        self.max_features = max_features
        self.embedding_dims = embedding_dims
        self.class_num = class_num
        self.last_cativation = last_cativation

    
    def get_model(self):
        input = Input((self.maxlen,))

        embedding = Embedding(self.max_features, self.embedding_dims, input_length=self.maxlen)(input)
        x = Bidirectional(LSTM(128, return_sequences=True))(embedding)
        x = Attention(self.maxlen)(x)
        
        x = Dense(128, activation='relu')(x)
        x = Dense(64, activation='relu')(x)
        x = Dense(32, activation='relu')(x)

        output = Dense(self.class_num, activation=self.last_cativation)(x)
        model = Model(inputs=input, outputs= output)
        model.summary()

        return model

3.2 数据处理

模型有了那就需要将数据处理为机器能够学习的，将所有字符替换为数字。【至于数据标注，就百度ai调一下，对数据进行标注就行。不放心就自己看一遍】

    txt_path = 'D:/weibo/dataprocess/data/data_label.txt'

    text_list, text_label = get_data(txt_path)
    train_data = []
    for text in text_list:
        train_data.append(build_chars(text))

    chars = []
    for word_list in train_data:
        for word in word_list:
            if word not in chars:
                chars.append(word)

    chars = pd.Series(chars).value_counts()
    chars[:] = range(1, len(chars) + 1)
    build_dict(chars)
    # for k, v in enumerate(chars):
    #     print('%d  %s' % (k, v))
    #     print(chars.index(v))

    maxlen = 64
    embedding_dim = 128

    train_x = []
    for data in train_data:
        temp = []
        for d in data:
            temp.append(chars[d])
        train_x.append(temp)

    train_x = sequence.pad_sequences(train_x, maxlen=maxlen, value=0, padding='post')
    train_y = text_label

3.3 训练

    train_x = np.array(train_x)
    # log.logger.debug(train_x)

    train_y = np.array(train_y)
    # log.logger.debug(train_y)

    textRnn = TextAttBiRNN(maxlen, len(chars) + 1, embedding_dim, class_num = 2)
    model = textRnn.get_model()
    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

    num_epochs = 5
    history = model.fit(
        train_x,
        train_y,
        epochs=num_epochs,
        verbose = 1
    )
    model.save_weights(model_path)

4. 结果

测试结果如下：

5.总结

这个项目做的很临时，基本采用的方式就是能实现。对于数据处理正常就行。
模型这里，采用的是我在天池新闻分类中使用的lstm+attention。感觉还行吧，作为基础练练手用。想要更香的方式就试一试bert。新闻分类有95%以上。lstm只有93%多一点。

最近其他事耽搁，更得很慢，请谅解。（有问题可以私信，看到会第一时间回复）

关注公众号回复关键词【情感分析模型】，即可下载模型构建代码。