摘要
gSpan算法是一种基于频繁图的数据挖掘算法。该算法基于无候选人产生的频繁子图，采用深度优先搜索策略挖掘频繁连接子图。由于其设计结构具有连续性以及无候选人产生，算法的性能得以提高，在执行速度上可以达到前人算法如FSG算法的15～100倍。基于化合物库Chemical_340测试发现，该算法能够以卓越性能有效挖掘频繁子图。该算法可以应用在搜索具有相同子结构的化合物研究中，对相关领域研究发展具有重要意义。
Introduces a graph - based substructure pattern mining algorithm called gSpan, which discovers frequent substructures without candidate generation and adopts the depth- first search strategy to mine frequent connected subgraphs efficiently. Its performance is enhanced because the continuous design and non - candidate. When it is applied on the chemical compound - library Chemical_ 340, gSpan substantially outperforms previous algorithms such as FSG, sometimes by an order of magnitude, gSpan can be applied in the research of finding compounds with same substructure, it is very important to related areas.
出处
《计算机技术与发展》
2007年第10期58-60,64,共4页
Computer Technology and Development
基金
国家科委创新基金资助项目(06C26213401229)
