本文将对Bipartite Graph正则方法进行总结。
1.Scalable Semi-Supervised Learning by Efficient Anchor Graph Regularization(2016KDE)
动机:AGR still has limitations in its two components: (1) in anchor graph construction, the estimation of the local weights between each datapoint and its neighboring anchors could be biased and relatively slow; and (2) in anchor graph regularization, the adjacency matrix that estimates the relationship between datapoints, is not sufficiently effective。
anchor点利用kmeans 计算,
假设f为标签预测函数
目标函数如下:
2.Fast Semisupervised Learning With Bipartite Graph for Large-Scale Data(2020TNN)
graph构造方法:
目标函数如下:
第一项为anchor图正则,第二项为计算得到的标签与原始标签的差异。
2.Fast Semi-Supervised Learning with Optimal Bipartite Graph(2020KED)
与文章1相比,多加了一项对anchor的优化。
3. Efficient and Robust MultiView Clustering With Anchor Graph Regularization(2022TCSVT)
目标函数如下:
第一项为anchor graph分解(熵范数替代二范数),第二项为图正则。
总结:
1.图正则的方法,利用anchor graph 构造 整体 graph,然后利用整体graph进行图正则;
2.anchor构造方法3种:高斯核kernel度量相似性,利用LLE 重构,利用文章2中的方法进行计算。
未完待续