Analysis via Orthonormal Systems in Reproducing Kernel Hilbert $C^{\ast }$-Modules and Applications

Abstract

Kernel methods have been among the most popular techniques in machine learning, where learning tasks are solved using the property of reproducing kernel Hilbert space (RKHS). In this paper, we propose a novel data analysis framework with reproducing kernel Hilbert $C^{\ast }$-module (RKHM), which is another generalization of RKHS than vector-valued RKHS (vv-RKHS). Analysis with RKHMs enables us to deal with structures among variables more explicitly than vv-RKHS. We show the theoretical validity for the construction of orthonormal systems in Hilbert $C^{\ast }$-modules, and derive concrete procedures for orthonormalization in RKHMs with those theoretical properties in numerical computations. Moreover, we apply those to generalize with RKHM kernel principal component analysis and the analysis of dynamical systems with Perron-Frobenius operators. The empirical performance of our methods is also investigated by using synthetic and real-world data.

References in corpus (9)

• Deep learning for universal linear embeddings of nonlinear dynamics
• Distinguishing cause from effect using observational data: methods and benchmarks
• Data-driven approximation of the Koopman generator: Model reduction, system identification, and control
• Subspace dynamic mode decomposition for stochastic Koopman analysis
• Operator-valued Kernels for Learning from Functional Response Data
• Dynamic mode decomposition in vector-valued reproducing kernel Hilbert spaces for extracting dynamical structure among observables
• Diagonalizing ''compact'' operators on Hilbert W*-modules
• On orthogonal systems in Hilbert C*-modules
• The Matrix Hilbert Space and Its Application to Matrix Learning

Cited by in corpus (1)

• Kernel Mean Embeddings of Von Neumann-Algebra-Valued Measures