10 May 2012 A computational approach for statistical learning and inference
Author Affiliations +
Abstract
In this paper, we demonstrate that a wide class of machine learning problems can be formulated as general problems of multi-valued decision and classification. To reduce the sample complexity associated with the statistical learning and inference schemes, we propose the principle of probabilistic comparison, the inclusion principle and exact computational methods for constructing multistage procedures for the relevant multi-hypothesis testing problems.
© (2012) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Xinjia Chen, "A computational approach for statistical learning and inference", Proc. SPIE 8401, Independent Component Analyses, Compressive Sampling, Wavelets, Neural Net, Biosystems, and Nanoengineering X, 84010T (10 May 2012); doi: 10.1117/12.918919; https://doi.org/10.1117/12.918919
PROCEEDINGS
12 PAGES


SHARE
KEYWORDS
Bismuth

Zinc

Machine learning

Stochastic processes

Evolutionary algorithms

Composites

Optimization (mathematics)

Back to Top