对于大文本进行挖掘聚类，该方法不考虑文字词语出现的频率信息，考虑上下文语境，将所有的字根据预定义的特征进行词位特征学习，获得一个训练模型。然后对待分字符串的每一个字进行词位标注，最后根据词位定义获得最终的分词结果。(Digging for large text clustering, the method does not consider the text word frequency of information, considering the context, all the words according to the characters of a predefined characteristics of word learning, to get a training model.Then each word of the character string is marked with the word bit, and finally the final word segmentation result is obtained according to the definition of the word bit.)