林绵数据挖掘Approximate Graph Mining with Label Costs.ppt

图挖掘可以粗略地分为频繁子图挖掘、图聚类与图分类等等.频繁子图挖掘的目的是找到在图集中频繁出现的子图 频繁子图挖掘结果可作为图聚类、图分类等其他图挖掘研究的基础,,频繁子图的挖掘结果已经应用于一些领域,如在毒理学预测项目中的应用,其目的是找到频繁出现的与有毒物质具有相同子结构的物质 * 衡量label之间的相似性 * * * 反单调性加快剪枝 * * Approximate Graph Mining with Label Costs Outline 2.圣诞节的习俗 3.圣诞老人 4.圣诞树 Introduction Preliminaries Computing Representative Sets Mining Algorithm Experimental Evaluation Conclusion Introduction Graphs are a natural way to model many of the modern complex datasets that typically have interlinked entities connected with various relationships. Our focus is on graph pattern discovery methods that can simult