迁移近邻传播聚类算法

Transfer Affinity Propagation Clustering Algorithm

在目标域可利用数据匮乏的场景下,传统聚类算法的性能往往会下降.在该场景下,通过抽取源域中的有用知识用于指导目标域学习以得到更为合适的类别信息和聚类性能,是一种有效的学习策略.借此提出一种基于近邻传播的迁移聚类(transfer affinity propagation,简称TAP)算法,在源域和目标域数据分布相似的情况下,通过引入迁移学习机制来改善近邻传播聚类(affinity propagation,简称AP)算法在数据匮乏场景下的聚类性能.为保证迁移的有效性,TAP在综合考虑源域和目标域的统计特性及几何特征的基础上改进AP算法中的消息传递机制使其具备迁移能力,从而达到辅助目标域学习的目的.此外,通过TAP对应的因子图,亦可说明TAP可以以类似AP的消息传递机制,在目标域数据匮乏的情况下进行高效的知识迁移,为最终所获得的聚类结果提供了保证.在模拟数据集和真实数据集上的仿真实验结果显示,所提出的算法较之经典AP算法在处理非充分数据聚类任务时具有更佳的性能.

The main limitation of most traditional clustering methods is that they cannot effectively deal with the insufficient datasets in target domain, It is desirable to develop new cluster algorithms which can leverage useful information in the source domain to guide the clustering performance in the target domain so that appropriate number of clusters and high quality clustering result can be obtained in this situation. In this paper, a clustering algorithm called transfer affinity propagation （TAP） is proposed for the insufficient dataset scenarios. The new algorithm improves the clustering performance when the distribution of source and target domains are similar. The basic idea of TAP is to modify the update rules about two message propagations, used in affinity propagation （AP）, through leveraging statistical property and geometric structure together. With the corresponding factor graph, TAP indeed can be applied to cluster in AP-like transfer learning, i.e., TAP earl abstract the knowledge of source domains through the two tricks to enhance the learning of target domain even if the data in the current scene are not adequate. Extensive experiments demonstrate that the proposed algorithm outperforms traditional algorithms in situations of insufficient data.