semantic matrix造句

It process documents not only based on latent semantic analysis , but also based on text multilevel dependency structure . the method first analysis the latent semantic structure of texts , make single value decomposition on text - matrix , reconstruct the semantic matrix ; then a method based on text multilevel dependency structure is adopted , deeply analysis the content of the semantic matrix , abstract the important sentences to generate abstraction and make up the shortage of latent semantic analysis on structure and syntax
首先通过对文本进行潜在语义分析,对文本矩阵进行相应的奇异值分解,重构语义矩阵;然后采用基于篇章多级依存结构的文摘分析方法,对重构的语义矩阵表示的文本内容进行深入的分析,抽取重要的句子生成文摘,这样就弥补了潜在语义分析在词法和句法分析上的不足;同时过滤和去除了语义噪音,缩小了问题的规模。
It's difficult to find semantic matrix in a sentence. 用semantic matrix造句挺难的