An empirical study of using an ensemble model in e-commerce taxonomy classification challenge

Yugang Jia; Xin Wang; Hanqing Cao; Boshu Ru; Tianzhong Yang

An empirical study of using an ensemble model in e-commerce taxonomy classification challenge

Yugang Jia, Xin Wang, Hanqing Cao, Boshu Ru, Tianzhong Yang

Research output: Contribution to journal › Conference article › peer-review

Abstract

In the Rakuten data challenge on taxonomy Classification for eCommerce - scale Product Catalogs, we propose an approach based on deep convolutional neural networks to predict product taxonomies using their descriptions. The classification performance of the proposed system is further improved with oversampling, threshold moving and error correct output coding. The best classification accuracy is obtained through ensembling multiple networks trained differently with multiple inputs comprising of various extracted features.

Original language	English (US)
Journal	CEUR Workshop Proceedings
Volume	2319
State	Published - 2018
Event	2018 SIGIR Workshop On eCommerce, eCom 2018 - Ann Arbor, United States Duration: Jul 12 2018 → …

Bibliographical note

Publisher Copyright:
Copyright © 2018 by the paper’s authors.

Keywords

Convolutional neural networks
Error correct output coding
Imbalanced classes
Multi-class classification
Word embedding

Find It

Find It at U of M Libraries

Cite this

@article{684839fbb52a441199806537b0a9ede4,

title = "An empirical study of using an ensemble model in e-commerce taxonomy classification challenge",

abstract = "In the Rakuten data challenge on taxonomy Classification for eCommerce - scale Product Catalogs, we propose an approach based on deep convolutional neural networks to predict product taxonomies using their descriptions. The classification performance of the proposed system is further improved with oversampling, threshold moving and error correct output coding. The best classification accuracy is obtained through ensembling multiple networks trained differently with multiple inputs comprising of various extracted features.",

keywords = "Convolutional neural networks, Error correct output coding, Imbalanced classes, Multi-class classification, Word embedding",

author = "Yugang Jia and Xin Wang and Hanqing Cao and Boshu Ru and Tianzhong Yang",

note = "Publisher Copyright: Copyright {\textcopyright} 2018 by the paper{\textquoteright}s authors.; 2018 SIGIR Workshop On eCommerce, eCom 2018 ; Conference date: 12-07-2018",

year = "2018",

language = "English (US)",

volume = "2319",

journal = "CEUR Workshop Proceedings",

issn = "1613-0073",

publisher = "CEUR-WS",

}

TY - JOUR

T1 - An empirical study of using an ensemble model in e-commerce taxonomy classification challenge

AU - Jia, Yugang

AU - Wang, Xin

AU - Cao, Hanqing

AU - Ru, Boshu

AU - Yang, Tianzhong

PY - 2018

Y1 - 2018

N2 - In the Rakuten data challenge on taxonomy Classification for eCommerce - scale Product Catalogs, we propose an approach based on deep convolutional neural networks to predict product taxonomies using their descriptions. The classification performance of the proposed system is further improved with oversampling, threshold moving and error correct output coding. The best classification accuracy is obtained through ensembling multiple networks trained differently with multiple inputs comprising of various extracted features.

AB - In the Rakuten data challenge on taxonomy Classification for eCommerce - scale Product Catalogs, we propose an approach based on deep convolutional neural networks to predict product taxonomies using their descriptions. The classification performance of the proposed system is further improved with oversampling, threshold moving and error correct output coding. The best classification accuracy is obtained through ensembling multiple networks trained differently with multiple inputs comprising of various extracted features.

KW - Convolutional neural networks

KW - Error correct output coding

KW - Imbalanced classes

KW - Multi-class classification

KW - Word embedding

UR - http://www.scopus.com/inward/record.url?scp=85062328694&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062328694&partnerID=8YFLogxK

M3 - Conference article

AN - SCOPUS:85062328694

SN - 1613-0073

VL - 2319

JO - CEUR Workshop Proceedings

JF - CEUR Workshop Proceedings

T2 - 2018 SIGIR Workshop On eCommerce, eCom 2018

Y2 - 12 July 2018

ER -

An empirical study of using an ensemble model in e-commerce taxonomy classification challenge

Abstract

Bibliographical note

Keywords

Find It

Other files and links

Fingerprint

Cite this