Kale-Linear

Knowledge-aware learning linear methods.

This package contains implementations of the following methods:

• Transformers (learning feature embeddings):
  • Multilinear Principal component analysis (MPCA), Lu et al., 2008, available at IEEE or Open Access.
  • Domain adaptation via transfer component analysis (TCA) [Pan et al., 2009].
  • Transfer Feature Learning with Joint Distribution Adaptation (JDA) [Long et al., 2013].
  • Balanced distribution adaptation (BDA) [Wang et al., 2017].
  • Maximum Independence Domain Adaptation (MIDA) [Yan et al., 2017].
• Estimators (learning classifiers):
  • Manifold Regularisation Learning Framework (LapSVM, LapRLS) [Belkin et al., 2006].
  • Adaptation Regularisation Learning Framework (ARSVM, ARRLS) [Long et al., 2014].
  • Covariate Independence Regularised Learning Framework (CoIRSVM, CoIRLS) [Zhou et al., 2020, Zhou, 2022].
  • Group-specific Discriminant Analysis (GSDA) [Zhou et al., 2025, Zhou, 2022].

Dependencies

• scikit-learn
• scipy
• numpy
• cvxopt
• osqp
• tensorly

Installation

Install the NumPy-based core package:

pip install kalelinear

Kale-Linear expects NumPy-compatible array inputs and returns NumPy arrays.

Other Transfer Learning/Domain Adaptation Repos on GitHub

• POT: Python Optimal Transport
• Everything about Transfer Learning
• ADA: (Yet) Another Domain Adaptation library
• Domain Adaptation & Transfer Learning Repositories
• Library of transfer learners and domain-adaptive classifiers
• domain-adaptation-toolbox
• Domain-Adaptations