随着互联网的发展与普及，在互联网中出现了越来越多不同的语言。用户在使用各类网络百科全书平台浏览文档时，如果想要找到内容相似的、用不同语言编写的文档，就需要通过词条之间的跨语言知识链接。但是当前不同语言编写的文档之间的跨语言知识链接还没有建立完全，需要通过分析文档之间的相似度进而判断是否可以在文档之间建立跨语言知识链接。而对于使用不同语言编辑的文档，目前多为人工判断其内容是否相似，这就导致了很高的人工成本。为了解决这个问题，论文将通过基于Siamese Neural Networks的LSTM（Long Short-Term Memory）模型，处理提取到的不同语言文档中的链接实体序列，自动化地判断文档之间的相似性，尝试对不同语言的文档进行匹配，并判断是否可以在不同语言文档之间建立起跨语言知识链接。

As the Internet is developing and prevailing, there emerge more and more languages on the Internet. When users want to find content-similar texts which are edited by different languages while looking through texts on the online encyclopedia, they will need to use cross-lingual knowledge linking. However, cross-lingual knowledge linking between texts edited by different languages is not established completely, and it requires analysis about similarity between texts edited by different languages. However, whether these texts edited by different languages have the same content is judged by human, and this can cost a lot. In order to solve this problem, this paper will evaluate the similarity between texts in different languages automatically by using LSTM (Long Short-Term Memory) Model which is built upon Siamese Neural Networks to process cross-lingual knowledge linking entities. By doing this, this paper will try to match texts in different languages if they are similar enough on the context and decide whether cross-lingual knowledge linking can be established between them.