Date Added: Jan 2012
The machine learning plays the key roles in many artificial intelligence areas including classification, regression and clustering. The traditional machine learning methods assume that the training and test sample are drawn from the same feature space and the same distribution. With the change of the distribution, the traditional machine learning methods need to rebuild the models using newly collected training samples. In real world, it is impossible or expensive to recollect and label the needed training samples and rebuild the models. To address the problem, the transfer learning is proposed.