跨语种文本检索是在不同语种文献之间实现检索。本文基于语境思想,提出了一种新的以文搜文的跨语种检索方法。本文方法将用户提交的检索文本转换成关键词共现关系的数组,再通过平行语料库的关键词共现关系映射分布模型映射成其他语种的关键词共现关系矩阵,再使用该矩阵实现其他语种文献的检索。该方法通过关键词的共现关系分布统计,有效地解决了词翻译歧义、词意模糊等问题,具有较好的检索准确率。 Cross-language text retrieval is to retrieve documents in different languages. Based on the contextual thinking, this paper proposes a new cross-language retrieval method for essay searching. This method transforms the retrieved texts submitted by users into an array of co-occurrences of keywords and then maps the co-occurrence relational matrices of keywords in other languages ​​through the co-occurrence relational mapping model of keyword co-occurrences in parallel corpus, and then uses the matrix to implement other languages Document retrieval. The method distributes the statistics of the co-occurrence relations of keywords and effectively solves the problems of word translation ambiguity and word meaning ambiguity, and has a good retrieval accuracy.