一种基于完全匹配和分词匹配的混合分词匹配算法

A hybrid participle approach based on the complete matching and participle matching

如何高效准确地进行酶的鉴别是代谢网络重构的难点和关键。由于基因标注文件中与ORF对应的蛋白质和酶信息都是采用自然语言描述的,重构中必须利用基因产物在酶数据库中寻找证据,从而进行与代谢相关的酶鉴别。针对目前通用的酶鉴别算法识别率低、效率低的问题,论文在对生物数据特性分析的基础上,提出了一种完全匹配和分词匹配相结合的混合分词匹配算法对酶信息进行鉴别,实验结果和分析表明,该算法具有更好的鉴别能力和更高的效率。

The enzyme identified is the most important part of identifying of metabolic activity and How to identify metabolic related enzymes is a special difficulty and a key step to make reconstruction network efficiency and accurately. Because of the character of the gene-annotate file, we must find evidence in the gene production by enzyme database to identify metabolic related enzymes. This paper proposes a hybrid participle algorithm which takes advantage of the feature that the problem＇ s feasible region is convex. Experimental and analytical results illustrate that the hybrid participle algorithm is more efficient and accurate.