本文介绍了我们在图上进行聚类算法研究的成果，特别是针对Graph Agglomerative Clustering (GAC) 的Graph Degree Linkage算法。该算法在性能上超越了谱聚类和归一化割，并且运行速度更快。简单易实现，无需依赖矩阵库。
Graph Agglomerative Clustering (GAC) toolbox
The distribution of Graph Agglomerative Clustering (GAC) toolbox is a summary of our work in the topic of agglomerative clustering on a graph.
Both graph-based clustering and agglomerative clustering have been studied extensively, while developing agglomerative clustering on a graph is a rarely investigated problem.
We proposed a simple and effective algorithm called graph degree linkage. It has better performance than normalized cuts and spectral clustering, and is much faster. In addition, the implementation of our algorithm is pretty simple. You don't need to have any matrix library (eigen-vector solver).
1542
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
