Scopus Indexed Publications

Paper Details


Title
Machine Learning and Data Analysis through Functional Analysis of Hilbert Spaces: Theory and Kernel-Based Applications

Author
, Md. Shafikul Islam,

Email

Abstract

This paper investigates and focuses on Kernel-based methods in Hilbert space theory for advancing machine learning and data analysis. They extend Euclidean spaces and also unitary spaces to infinite dimensions such as inner products, orthogonality, projections, the polarization identity, and convergence, to be applied to complex data. The study emphasizes the connection between functional analytic tools including operator theory, classical inequalities, and inner product identities, and modern machine learning models including support vector machines, Gaussian process regression, and deep kernel methods. These properties are central to the development of reproducing kernel Hilbert spaces (RKHS), which enable efficient nonlinear learning and underpin widely used algorithms. The practical relevance of Hilbert spaces is demonstrated through their role in machine learning algorithms such as Support Vector Machines (SVM), Principal Component Analysis (PCA), and kernel methods using Reproducing Kernel Hilbert Spaces (RKHS). Numerical simulations and MATLAB visualizations are employed to aid understanding and demonstrate the application of inner product theory in data-driven tasks such as classification and dimensionality reduction. The results show how the geometry of Hilbert spaces naturally supports core operations in machine learning, making them indispensable in theoretical development and algorithmic design. This flexibility allows it to model highly complex decision boundaries. And the polynomial kernel maps data into a polynomial feature space. This connection highlights the deep relationship between functional analysis and practical machine learning.


Keywords
Hilbert Space, RKHS, Kernel Methods, SVM, PCA

Journal or Conference Name
Mathematics and Statistics

Publication Year
2026

Indexing
scopus