Research on Partial Label Learning Algorithm Based on Self-pacedandTransfer Learning
In partial label learning,each training sample corresponds to a set of candidate labels,and only one label is a real label.Almost all existing partial label learning algorithms attempt to eliminate the ambiguity of the candidate label set by processing all labels indiscriminately.However,this approach lacks consideration of the complexity between labels and instances during training.In order to solve this limitation,this paper introduces self-paced learning and transfer learning to solve the partial label learning problem,and constructs a classifier with more stability and classification performance to solve the partial label problem in a novel way.
partial label learningself-paced learningtransfer learningalgorithm research
NETL
NSTL
万方数据
