基于FSG的最大频繁子图挖掘算法*

图挖掘已成为数据挖掘领域研究的热点，然而挖掘全部频繁子图很困难且得到的频繁子图过多，影响结果的理解和应用。可通过挖掘最大频繁子图来解决挖掘结果数量巨大的问题，最大频繁子图挖掘得到的结果数量很少且不丢失信息，节省了空间和以后的分析工作。基于算法FSG提出了最大频繁子图挖掘算法FSG-MaxGraph;结合节点的度、标记及邻接列表来计算规范编码，提出两个定理来减少子图同构判断的次数，并应用改进后的决策树来计算支持度。实验证明,新算法解决了挖掘结果太多理解困难的问题，且提高了挖掘效率。

FSG-based algorithm for mining maximal frequent subgraph

Graph mining has become a hot topic in the field of data mining, however, mining all frequent subgraph is very difficult and will get excessive frequent subgraph this impact on the understanding and application of the outcome. Through mi-ning maximal frequent subgraph to solve the problems of the number of the result is huge. Maximal frequent subgraph mining obtained a small number of the results and this without loss information, mining maximal frequent subgraph saved space and the work of analysis. This paper based on algorithm FSG proposed an algorithm FSG-MaxGraph for mining maximal frequent subgraphs. Combined the degree, nodes and adjacency list to calculating normal matrix coding and proposed two theorems that could reduce the times of subgraph isomorphism this improve the efficiency of the algorithm. Last, used the improved decision tree to computing support. The experiment can prove the new algorithm can solve the problem of the mining results difficult to understand and this new algorithm can improve the efficiency of mining.