Kernel Density Machines
This work addresses a foundational statistical estimation problem with potential applications in machine learning and data analysis, though it appears incremental as it builds on existing kernel methods.
The authors tackled the problem of estimating Radon-Nikodym derivatives in nonparametric settings by introducing kernel density machines (KDM), which achieved computational efficiency with low-rank approximations and provided rigorous theoretical guarantees like asymptotic consistency and finite-sample error bounds.
We introduce kernel density machines (KDM), a nonparametric estimator of a Radon--Nikodym derivative, based on reproducing kernel Hilbert spaces. KDM applies to general probability measures on countably generated measurable spaces under minimal assumptions. For computational efficiency, we incorporate a low-rank approximation with precisely controlled error that grants scalability to large-sample settings. We provide rigorous theoretical guarantees, including asymptotic consistency, a functional central limit theorem, and finite-sample error bounds, establishing a strong foundation for practical use. Empirical results based on simulated and real data demonstrate the efficacy and precision of KDM.