Using image-based and text-based information for sales prediction: A deep neural network model

Ying Wang, Yue Guo, Jaeki Song

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper is to investigate how text-based and image-based information influence product sales in electronic markets. We apply signaling theory to elaborate the role of image-based and text-based information in consumers' purchase decisions and use deep neural networks model to analyze different types of information in online sales websites. We collect information about 4, 368 furniture products from Amazon and find that both image-based and text-based information influence consumers' purchase decisions, but the former one is more crucial. This paper makes contributions to e-commerce literature by elaborating the signaling role of available information in sales websites, highlighting the importance of considering both text-based and image-based information in the data analysis, and demonstrating how to apply advanced deep learning techniques and models in e-commerce studies.

Original languageEnglish
Title of host publicationAmericas Conference on Information Systems 2018
Subtitle of host publicationDigital Disruption, AMCIS 2018
PublisherAssociation for Information Systems
ISBN (Print)9780996683166
StatePublished - 2018
Event24th Americas Conference on Information Systems 2018: Digital Disruption, AMCIS 2018 - New Orleans, United States
Duration: Aug 16 2018Aug 18 2018

Publication series

NameAmericas Conference on Information Systems 2018: Digital Disruption, AMCIS 2018

Conference

Conference24th Americas Conference on Information Systems 2018: Digital Disruption, AMCIS 2018
CountryUnited States
CityNew Orleans
Period08/16/1808/18/18

Keywords

  • Deep learning
  • Image
  • Neural network model
  • Signaling theory
  • Text
  • Unstructured data

Fingerprint Dive into the research topics of 'Using image-based and text-based information for sales prediction: A deep neural network model'. Together they form a unique fingerprint.

Cite this