word2vec算法tensorflow实现

word2vec有两种模型，一个是CBOW模型，另一个是Skip-gram模型。那么这里主要用tensorflow实现Skip-gram模型。

Skip-gram模型是在已知目标词的前提下预测其上下文。那么skip-gram模型只用到了一个隐藏层，一个输出层，输入及输出标注为 初始化的词向量。

那么原理如下：

1、设定固定的窗口大小来遍历训练文本，构建（中心词，临近词）的关系，使用网络模型来拟合两者的关系。

2、输入为中心词初始词向量，输出标注为临近词的初始词向量，通过网络模型利用输入与输出两组词向量来构建损失函数，隐藏层不使用激活函数，输出层结果为概率分数，使用softmax归一化。

3、最终产生的词向量是训练得到的隐藏层参数，不使用最后的输出。隐藏层节点个数即最终词向量空间的维度。

这篇博客讲解的很清楚，可以参考https://blog.csdn.net/rlnlo2pnefx9c/article/details/78747970，接下来主要是实现了这部分博客里面的思想。

那么接下来将要使用tensorflow实现Skip-gram模型，主要使用了python深度学习这本书中的代码。

下面函数代码的功能是读取text9压缩文件的内容，然后把其保存在words中，并显示了前面10个单词，运行结果如下图。

def read_data(filename):
    with zipfile.ZipFile(filename) as f:
        data = tf.compat.as_str(f.read(f.namelist()[0])).split()
    return data

words = read_data("text9.zip")
print(words[:10])
print('Data size %d' % len(words))

下面代码主要是建立字典，就是单词索引号的映射，一个单词对应一个索引号，并保存到dictionary中。count里面记录了每个单词出现的次数并把出现次数最多的前vocabulary_size-1个原素添加到count中。运行结果如下图所示：

vocabulary_size = 20
def build_dataset(words):
    count = [['UNK',-1]]
    count.extend(collections.Counter(words).most_common(vocabulary_size-1))
    dictionary = dict()  #创建空字典
    for word,_ in count:
        dictionary[word] = len(dictionary)
    print('count是:',end=' ')
    print(count)
    print("dictionary是:",end=' ')
    print(dictionary)    #dictionary 是  值:索引号
    data = list()
    unk_count = 0 #记录了不在字典中的词的个数
    for word in words:
        if word in dictionary:
            index = dictionary[word]  #如果该词在字典中出现，那么就保存在字典中对应的索引号
        else:
            index = 0        #否则就没有在字典中出现，默认为0
            unk_count = unk_count + 1
        data.append(index)
    count[0][1] = unk_count
    reverse_dictionary = dict(zip(dictionary.values(), dictionary.keys())) #reverse_dictionary是  索引号:值
    print("reverse_dictionary是:",end=' ')
    print(reverse_dictionary)
    return data, count, dictionary, reverse_dictionary


data, count, dictionary, reverse_dictionary = build_dataset(words)
print('Most common words (+UNK)', count[:5])
print('Sample data', data[:10])
del words  # Hint to reduce memory.

下面这个代码段主要就是获取上下文窗口，并且需要调整步数以及窗口值的大小。那么运行结果如下图所示：

data_index = 0

def generate_batch(batch_size, num_skips, skip_window):
    global data_index
    assert batch_size % num_skips == 0
    assert num_skips <= 2 * skip_window
    batch = np.ndarray(shape=(batch_size), dtype=np.int32)
    labels = np.ndarray(shape=(batch_size, 1), dtype=np.int32)
    span = 2 * skip_window + 1  # [ skip_window target skip_window ]
    buffer = collections.deque(maxlen=span)
    for _ in range(span):
        buffer.append(data[data_index])
        data_index = (data_index + 1) % len(data)
    for i in range(batch_size // num_skips):
        target = skip_window  # target label at the center of the buffer
        targets_to_avoid = [skip_window]
        for j in range(num_skips):
            while target in targets_to_avoid:
                target = random.randint(0, span - 1)
            targets_to_avoid.append(target)
            batch[i * num_skips + j] = buffer[skip_window]
            labels[i * num_skips + j, 0] = buffer[target]
        buffer.append(data[data_index])
        data_index = (data_index + 1) % len(data)
    return batch, labels

print('data:', [reverse_dictionary[di] for di in data[:8]])

for num_skips, skip_window in [(2, 1), (4, 2)]:
    data_index = 0
    batch, labels = generate_batch(batch_size=8, num_skips=num_skips, skip_window=skip_window)
    print('\nwith num_skips = %d and skip_window = %d:' % (num_skips, skip_window))
    print('    batch:', [reverse_dictionary[bi] for bi in batch])
    print('    labels:', [reverse_dictionary[li] for li in labels.reshape(8)])

那么下面主要是进行神经网络模型的训练，然后优化这个代价函数，得到隐藏层的权重值。

batch_size = 8
embedding_size = 8 # Dimension of the embedding vector.
skip_window = 1 # How many words to consider left and right.
num_skips = 2 # How many times to reuse an input to generate a label.
#我们选择一个随机验证集来对最近的邻居进行采样
#在这里，我们将验证样本限制为具有较低数字ID的词
valid_size = 2 # Random set of words to evaluate similarity on.
valid_window = 6 # Only pick dev samples in the head of the distribution.
valid_examples = np.array(random.sample(range(valid_window), valid_size))
num_sampled = 4 # Number of negative examples to sample.

graph = tf.Graph()

with graph.as_default(), tf.device('/cpu:0'):

    # Input data.
    train_dataset = tf.placeholder(tf.int32, shape=[batch_size])
    train_labels = tf.placeholder(tf.int32, shape=[batch_size, 1])
    valid_dataset = tf.constant(valid_examples, dtype=tf.int32)

    # Variables.
    embeddings = tf.Variable(
        tf.random_uniform([vocabulary_size, embedding_size], -1.0, 1.0))
    softmax_weights = tf.Variable(
        tf.truncated_normal([vocabulary_size, embedding_size],
                            stddev=1.0 / math.sqrt(embedding_size)))
    softmax_biases = tf.Variable(tf.zeros([vocabulary_size]))

    # Model.
    # Look up embeddings for inputs.
    embed = tf.nn.embedding_lookup(embeddings, train_dataset)
    # Compute the softmax loss, using a sample of the negative labels each time.
    loss = tf.reduce_mean(
        tf.nn.sampled_softmax_loss(weights=softmax_weights, biases=softmax_biases, inputs=embed,
                                   labels=train_labels, num_sampled=num_sampled, num_classes=vocabulary_size))

    # Optimizer.
    # ＃注意：优化器将优化softmax_weights和嵌入。
    # ＃这是因为嵌入被定义为一个可变的数量和。
    # ＃优化器的`minim`方法默认会修改所有变量的数量。
    # ＃这有助于张量传递。
    # ＃查看关于`tf.train.Optimizer.minimize（）`的文档了解更多细节

    optimizer = tf.train.AdagradOptimizer(1.0).minimize(loss)

    # 计算minibatch示例和所有嵌入之间的相似度。
    # 我们使用余弦距离：

    norm = tf.sqrt(tf.reduce_sum(tf.square(embeddings), 1, keepdims=True))
    normalized_embeddings = embeddings / norm
    valid_embeddings = tf.nn.embedding_lookup(
        normalized_embeddings, valid_dataset)
    similarity = tf.matmul(valid_embeddings, tf.transpose(normalized_embeddings))

 下面代码是启动这个会话，然后激活变量，打印在运行过程中损失值的变化情况。

num_steps = 10
with tf.Session(graph=graph) as session:
    tf.global_variables_initializer().run()
    print('Initialized')
    average_loss = 0
    for step in range(num_steps):
        batch_data, batch_labels = generate_batch(
            batch_size, num_skips, skip_window)
        feed_dict = {train_dataset: batch_data, train_labels: batch_labels}
        _, l = session.run([optimizer, loss], feed_dict=feed_dict)
        average_loss += l
        if step % 1 == 0:
            if step > 0:
                average_loss = average_loss / 2000
            # The average loss is an estimate of the loss over the last 2000 batches.
            print('Average loss at step %d: %f' % (step, average_loss))
            average_loss = 0
        # note that this is expensive (~20% slowdown if computed every 500 steps)
        if step % 10 == 0:
            sim = similarity.eval()
            for i in range(valid_size):
                valid_word = reverse_dictionary[valid_examples[i]]
                top_k = 8  # number of nearest neighbors
                nearest = (-sim[i, :]).argsort()[1:top_k + 1]
                log = 'Nearest to %s:' % valid_word
                for k in range(top_k):
                    close_word = reverse_dictionary[nearest[k]]
                    log = '%s %s,' % (log, close_word)
                print(log)
    final_embeddings = normalized_embeddings.eval()

下面这个代码主要是用于可视化这个结果 。如下图所示：

num_points = 19

tsne = TSNE(perplexity=30, n_components=2, init='pca', n_iter=5000)
two_d_embeddings = tsne.fit_transform(final_embeddings[1:num_points+1, :])


def plot(embeddings, labels):
    assert embeddings.shape[0] >= len(labels), 'More labels than embeddings'
    pylab.figure(figsize=(15, 15))  # in inches
    for i, label in enumerate(labels):
        x, y = embeddings[i, :]
        pylab.scatter(x, y)
        pylab.annotate(label, xy=(x, y), xytext=(5, 2), textcoords='offset points',
                       ha='right', va='bottom')
    pylab.show()


words = [reverse_dictionary[i] for i in range(1, num_points + 1)]
plot(two_d_embeddings, words)