GLD

GLD refers to generalized label distribution (learning). It extends LDL by allowing the supervised target to be a generalized label distribution whose components are not required to lie on the probability simplex, as described in [GLD-WLYJ26].

BaseGLD

class pyldl.algorithms.base.BaseGLD(**kwargs)

Base class for all GLD models in PyLDL.

GLD-SVR

class pyldl.algorithms.GLD_SVR(**kwargs)

GLD-SVR is proposed in paper [GLD-WLYJ26].

GLD-\(k\)NN

class pyldl.algorithms.GLD_KNN(*, k: int = 5, **kwargs)

GLD-kNN is proposed in paper [GLD-WLYJ26].

GLD-BFGS

class pyldl.algorithms.GLD_BFGS(**kwargs)

GLD-BFGS is proposed in paper [GLD-WLYJ26].

References

[GLD-WLYJ26] (1,2,3,4)

Haitao Wu, Weiwei Li, Kun Yue, and Xiuyi Jia. Learning generalized label distributions. In Proceedings of the International Conference on Machine Learning. 2026. Early Access.