Cluster and dynamic-TrAdaBoost-based transfer learning for text classification
Zibin Li, Bo Liu, Yanshan Xiao · 2017
Transfer learning can use the knowledge of closed fields to enhance complete the learning tasks in the target field. After several years of development, it has been widely used in research, data mining and deep learning. The problem that the sample weight of TrAdaBoost is prone to polarization in the source domain and the target domain when is in the transfer learning process. So we propose a transfer learning algorithm which based on dynamic-TrAdaBoost and k-means algorithm. This experiment proved to be a good solution to this problem.