| Summary: | 碩士 === 國立東華大學 === 資訊工程學系 === 100 === Due to the di±culty of studying a huge graph, graph sampling issue is in great demand. According to di®erent requirements of the samples, we can use di®erent methods to solve the graph sampling problem. In this thesis, we propose two algo-rithms. Our method can be divided into two parts, namely, edge reduction and vertex reduction. In the edge reduction, we try to keep the shape of the reduction graph as similar as the original one by reducing some edges in dense regions. It can avoid too many changes of the graph. In vertex reduction, we preserve the most important vertices such that they can still play important roles in the reduction graph. Some experimental results show that our methods are practical.
|
|---|
