Minimizing response time in time series classification

Shin Ando, Einoshin Suzuki

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Providing a timely output is one of the important criteria in applications of time series classification. Recent studies have been motivated to explore models of early prediction, prediction based on truncated temporal observations. The truncation of input improves the response time, but generally reduces the reliability of the prediction. The trade-off between the earliness and the accuracy is an inherent challenge of learning an early prediction model. In this paper, we present an optimization-based approach for learning an ensemble model for timely prediction with an intuitive objective function. The proposed model is comprised of time series classifiers with different response time, and a sequential aggregation procedure to determine the single timing of its output. We formalize the training of the ensemble classifier as a quadratic programming problem and present an iterative algorithm which minimizes an empirical risk function and the response time required to achieve the minimal risk simultaneously. We conduct an empirical study using a collection of behavior and time series datasets to evaluate the proposed algorithm. In the comparisons of the traditional and time-sensitive performance measures, the ensemble framework showed significant advantages over the existing methods on early prediction.

Original languageEnglish
Pages (from-to)449-476
Number of pages28
JournalKnowledge and Information Systems
Volume46
Issue number2
DOIs
Publication statusPublished - Feb 1 2016

Fingerprint

Time series
Classifiers
Quadratic programming
Agglomeration

All Science Journal Classification (ASJC) codes

  • Software
  • Information Systems
  • Human-Computer Interaction
  • Hardware and Architecture
  • Artificial Intelligence

Cite this

Minimizing response time in time series classification. / Ando, Shin; Suzuki, Einoshin.

In: Knowledge and Information Systems, Vol. 46, No. 2, 01.02.2016, p. 449-476.

Research output: Contribution to journalArticle

@article{78b461a0252047d29d889c35996f26b1,
title = "Minimizing response time in time series classification",
abstract = "Providing a timely output is one of the important criteria in applications of time series classification. Recent studies have been motivated to explore models of early prediction, prediction based on truncated temporal observations. The truncation of input improves the response time, but generally reduces the reliability of the prediction. The trade-off between the earliness and the accuracy is an inherent challenge of learning an early prediction model. In this paper, we present an optimization-based approach for learning an ensemble model for timely prediction with an intuitive objective function. The proposed model is comprised of time series classifiers with different response time, and a sequential aggregation procedure to determine the single timing of its output. We formalize the training of the ensemble classifier as a quadratic programming problem and present an iterative algorithm which minimizes an empirical risk function and the response time required to achieve the minimal risk simultaneously. We conduct an empirical study using a collection of behavior and time series datasets to evaluate the proposed algorithm. In the comparisons of the traditional and time-sensitive performance measures, the ensemble framework showed significant advantages over the existing methods on early prediction.",
author = "Shin Ando and Einoshin Suzuki",
year = "2016",
month = "2",
day = "1",
doi = "10.1007/s10115-015-0826-7",
language = "English",
volume = "46",
pages = "449--476",
journal = "Knowledge and Information Systems",
issn = "0219-1377",
publisher = "Springer London",
number = "2",

}

TY - JOUR

T1 - Minimizing response time in time series classification

AU - Ando, Shin

AU - Suzuki, Einoshin

PY - 2016/2/1

Y1 - 2016/2/1

N2 - Providing a timely output is one of the important criteria in applications of time series classification. Recent studies have been motivated to explore models of early prediction, prediction based on truncated temporal observations. The truncation of input improves the response time, but generally reduces the reliability of the prediction. The trade-off between the earliness and the accuracy is an inherent challenge of learning an early prediction model. In this paper, we present an optimization-based approach for learning an ensemble model for timely prediction with an intuitive objective function. The proposed model is comprised of time series classifiers with different response time, and a sequential aggregation procedure to determine the single timing of its output. We formalize the training of the ensemble classifier as a quadratic programming problem and present an iterative algorithm which minimizes an empirical risk function and the response time required to achieve the minimal risk simultaneously. We conduct an empirical study using a collection of behavior and time series datasets to evaluate the proposed algorithm. In the comparisons of the traditional and time-sensitive performance measures, the ensemble framework showed significant advantages over the existing methods on early prediction.

AB - Providing a timely output is one of the important criteria in applications of time series classification. Recent studies have been motivated to explore models of early prediction, prediction based on truncated temporal observations. The truncation of input improves the response time, but generally reduces the reliability of the prediction. The trade-off between the earliness and the accuracy is an inherent challenge of learning an early prediction model. In this paper, we present an optimization-based approach for learning an ensemble model for timely prediction with an intuitive objective function. The proposed model is comprised of time series classifiers with different response time, and a sequential aggregation procedure to determine the single timing of its output. We formalize the training of the ensemble classifier as a quadratic programming problem and present an iterative algorithm which minimizes an empirical risk function and the response time required to achieve the minimal risk simultaneously. We conduct an empirical study using a collection of behavior and time series datasets to evaluate the proposed algorithm. In the comparisons of the traditional and time-sensitive performance measures, the ensemble framework showed significant advantages over the existing methods on early prediction.

UR - http://www.scopus.com/inward/record.url?scp=84958152677&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958152677&partnerID=8YFLogxK

U2 - 10.1007/s10115-015-0826-7

DO - 10.1007/s10115-015-0826-7

M3 - Article

AN - SCOPUS:84958152677

VL - 46

SP - 449

EP - 476

JO - Knowledge and Information Systems

JF - Knowledge and Information Systems

SN - 0219-1377

IS - 2

ER -