Competitive Online Generalised Linear Regression with Multidimensional Outputs. / Dzhamtyrova, Raisa; Kalnishkan, Yury.

Proceedings of IJCNN 2019. IEEE, 2019. p. 1-8.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Published

Standard

Competitive Online Generalised Linear Regression with Multidimensional Outputs. / Dzhamtyrova, Raisa; Kalnishkan, Yury.

Proceedings of IJCNN 2019. IEEE, 2019. p. 1-8.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Harvard

APA

Vancouver

Author

BibTeX

@inproceedings{7fab20e5e3244158bec1371ee3d8e919,
title = "Competitive Online Generalised Linear Regression with Multidimensional Outputs",
abstract = "We apply online prediction with expert advice to construct a universal algorithm for multi-class classification problem. Our experts are generalised linear regression models with multidimensional outputs, i.e. neural networks with multiple output nodes but no hidden nodes. We allow the final layer transfer function to be a softmax function with linear activations to all output neurons. We build an online algorithm competitive with all the experts of relevant models of this type and derive an upper bound on the cumulative loss of the algorithm. We carry out experiments on three data sets and compare cumulative losses of our algorithm and a single neuron with multiple output nodes.",
keywords = "on-line learning, regression, kullback-leibler distance, competitive prediction",
author = "Raisa Dzhamtyrova and Yury Kalnishkan",
year = "2019",
doi = "10.1109/IJCNN.2019.8851941",
language = "English",
isbn = "978-1-7281-1986-1",
pages = "1--8",
booktitle = "Proceedings of IJCNN 2019",
publisher = "IEEE",

}

RIS

TY - GEN

T1 - Competitive Online Generalised Linear Regression with Multidimensional Outputs

AU - Dzhamtyrova, Raisa

AU - Kalnishkan, Yury

PY - 2019

Y1 - 2019

N2 - We apply online prediction with expert advice to construct a universal algorithm for multi-class classification problem. Our experts are generalised linear regression models with multidimensional outputs, i.e. neural networks with multiple output nodes but no hidden nodes. We allow the final layer transfer function to be a softmax function with linear activations to all output neurons. We build an online algorithm competitive with all the experts of relevant models of this type and derive an upper bound on the cumulative loss of the algorithm. We carry out experiments on three data sets and compare cumulative losses of our algorithm and a single neuron with multiple output nodes.

AB - We apply online prediction with expert advice to construct a universal algorithm for multi-class classification problem. Our experts are generalised linear regression models with multidimensional outputs, i.e. neural networks with multiple output nodes but no hidden nodes. We allow the final layer transfer function to be a softmax function with linear activations to all output neurons. We build an online algorithm competitive with all the experts of relevant models of this type and derive an upper bound on the cumulative loss of the algorithm. We carry out experiments on three data sets and compare cumulative losses of our algorithm and a single neuron with multiple output nodes.

KW - on-line learning

KW - regression

KW - kullback-leibler distance

KW - competitive prediction

U2 - 10.1109/IJCNN.2019.8851941

DO - 10.1109/IJCNN.2019.8851941

M3 - Conference contribution

SN - 978-1-7281-1986-1

SP - 1

EP - 8

BT - Proceedings of IJCNN 2019

PB - IEEE

ER -