Stream-suitable optimization algorithms for some soft-margin support vector machine variants

Soft-margin support vector machines (SVMs) are an important class of classification models that are well known to be highly accurate in a variety of settings and over many applications. The training of SVMs usually requires that the data be available all at once, in batch. The Stochastic majorizatio...

Full description

Saved in:
Bibliographic Details
Published in:Japanese journal of statistics and data science Vol. 1; no. 1; pp. 81 - 108
Main Authors: Nguyen, Hien D., Jones, Andrew T., McLachlan, Geoffrey J.
Format: Journal Article
Language:English
Published: Singapore Springer Singapore 01.06.2018
Subjects:
Chemistry and Earth Sciences
Computer Science
Economics
Finance
Health Sciences
Humanities
Insurance
Law
Management
Mathematics and Statistics
Medicine
Physics
Statistical Theory and Methods
Statistics
Statistics and Computing/Statistics Programs
Statistics for Business
Statistics for Engineering
Statistics for Life Sciences
Statistics for Social Sciences
Support vector machines
MNIST
Big data
Stochastic majorization–minimization algorithm
Streamed data
ISSN:2520-8756, 2520-8764
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Soft-margin support vector machines (SVMs) are an important class of classification models that are well known to be highly accurate in a variety of settings and over many applications. The training of SVMs usually requires that the data be available all at once, in batch. The Stochastic majorization–minimization (SMM) algorithm framework allows for the training of SVMs on streamed data instead. We utilize the SMM framework to construct algorithms for training hinge loss, squared-hinge loss, and logistic loss SVMs. We prove that our three SMM algorithms are each convergent and demonstrate that the algorithms are comparable to some state-of-the-art SVM-training methods. An application to the famous MNIST data set is used to demonstrate the potential of our algorithms.
ISSN:2520-8756
2520-8764
DOI:10.1007/s42081-018-0001-y