🤝🤝🤝 Please star ⭐️ it for promoting open source projects 🌍 ! Thanks !

A transfer learning regression model based on Kernel Mean Matching (KMM) algorithm

Written using Python, which is suitable for operating systems, e.g., Windows/Linux/MAC OS etc.

pip install KMMTR 

pip show KMMTR 

pip install --upgrade KMMTR

Click Here

Huang, J., Gretton, A., Borgwardt, K., Schölkopf, B., & Smola, A. (2006). Correcting sample selection bias by unlabeled data. Advances in neural information processing systems, 19.

Maintained by Bin Cao. Please feel free to open issues in the Github or contact Bin Cao (bcao@shu.edu.cn) in case of any problems/comments/suggestions in using the code.

1 : Instance-based transfer learning

• Instance selection (marginal distributions are same while conditional distributions are different) :

  TrAdaboost

• Instance re-weighting (conditional distributions are same while marginal distributions are different) :

  KMM

2 : Feature-based transfer learning

• Explicit distance:

  • case 1 : marginal distributions are same while conditional distributions are different:

    TCA(MMD based) ; DAN(MK-MMD based)

  • case 1 : conditional distributions are same while marginal distributions are different

    JDA

  • case 3 : Both marginal distributions and conditional distributions are different

    DDA

• Implicit distance :

  DANN

3 : Parameter-based transfer learning

• Pretraining + fine tune