Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning

In this paper, we investigate the basic properties of binary classification with a pseudo model based on the Itakura–Saito distance and reveal that the Itakura–Saito distance is a unique appropriate measure for estimation with the pseudo model in the framework of general Bregman divergence. Furtherm...

Full description

Bibliographic Details
Main Authors: Takashi Takenouchi, Osamu Komori, Shinto Eguchi
Format: Article
Language:English
Published: MDPI AG 2015-08-01
Series:Entropy
Subjects:
Online Access:http://www.mdpi.com/1099-4300/17/8/5673
id doaj-41c4a7329be84343b7f879df451a242c
record_format Article
spelling doaj-41c4a7329be84343b7f879df451a242c2020-11-25T00:30:24ZengMDPI AGEntropy1099-43002015-08-011785673569410.3390/e17085673e17085673Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task LearningTakashi Takenouchi0Osamu Komori1Shinto Eguchi2Future University Hakodate, 116-2 Kamedanakano, Hakodate Hokkaido 041-8655, JapanThe Institute of Statistical Mathematics, 10-3 Midori-cho, Tachikawa, Tokyo 190-8562, JapanThe Institute of Statistical Mathematics, 10-3 Midori-cho, Tachikawa, Tokyo 190-8562, JapanIn this paper, we investigate the basic properties of binary classification with a pseudo model based on the Itakura–Saito distance and reveal that the Itakura–Saito distance is a unique appropriate measure for estimation with the pseudo model in the framework of general Bregman divergence. Furthermore, we propose a novelmulti-task learning algorithm based on the pseudo model in the framework of the ensemble learning method. We focus on a specific setting of the multi-task learning for binary classification problems. The set of features is assumed to be common among all tasks, which are our targets of performance improvement. We consider a situation where the shared structures among the dataset are represented by divergence between underlying distributions associated with multiple tasks. We discuss statistical properties of the proposed method and investigate the validity of the proposed method with numerical experiments.http://www.mdpi.com/1099-4300/17/8/5673multi-task learningItakura–Saito distancepseudo modelun-normalized model
collection DOAJ
language English
format Article
sources DOAJ
author Takashi Takenouchi
Osamu Komori
Shinto Eguchi
spellingShingle Takashi Takenouchi
Osamu Komori
Shinto Eguchi
Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
Entropy
multi-task learning
Itakura–Saito distance
pseudo model
un-normalized model
author_facet Takashi Takenouchi
Osamu Komori
Shinto Eguchi
author_sort Takashi Takenouchi
title Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
title_short Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
title_full Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
title_fullStr Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
title_full_unstemmed Binary Classification with a Pseudo Exponential Model and Its Application for Multi-Task Learning
title_sort binary classification with a pseudo exponential model and its application for multi-task learning
publisher MDPI AG
series Entropy
issn 1099-4300
publishDate 2015-08-01
description In this paper, we investigate the basic properties of binary classification with a pseudo model based on the Itakura–Saito distance and reveal that the Itakura–Saito distance is a unique appropriate measure for estimation with the pseudo model in the framework of general Bregman divergence. Furthermore, we propose a novelmulti-task learning algorithm based on the pseudo model in the framework of the ensemble learning method. We focus on a specific setting of the multi-task learning for binary classification problems. The set of features is assumed to be common among all tasks, which are our targets of performance improvement. We consider a situation where the shared structures among the dataset are represented by divergence between underlying distributions associated with multiple tasks. We discuss statistical properties of the proposed method and investigate the validity of the proposed method with numerical experiments.
topic multi-task learning
Itakura–Saito distance
pseudo model
un-normalized model
url http://www.mdpi.com/1099-4300/17/8/5673
work_keys_str_mv AT takashitakenouchi binaryclassificationwithapseudoexponentialmodelanditsapplicationformultitasklearning
AT osamukomori binaryclassificationwithapseudoexponentialmodelanditsapplicationformultitasklearning
AT shintoeguchi binaryclassificationwithapseudoexponentialmodelanditsapplicationformultitasklearning
_version_ 1725326843508162560