word2vec本质上来说就是一个矩阵分解的模型，简单地说，矩阵刻画了每个词和其上下文的词的集合的相关情况。对这个矩阵进行分解，只取每个词对应在隐含空间的向量。(Word2vec is essentially a model of matrix decomposition. In a simple way, matrices characterize the set of words in each word and its context. The matrix is decomposed,