TY - JOUR
T1 - Image or text
T2 - Which one is more influential? a deep-learning approach for visual and textual data analysis in the digital economy
AU - Wang, Ying
AU - Song, Jaeki
N1 - Publisher Copyright:
© 2020 by the Association for Information Systems.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2020
Y1 - 2020
N2 - In a digital economy, different types of information about products communicate their quality and characteristics to prospective consumers. However, it remains unclear which type of information plays the most important role in individuals’ decision-making processes. In this study, we explore the effect that unstructured data has on and the importance of congruence between textual and visual data in consumers’ purchase decisions. We apply a deep neural network model to rank the importance of different information types and use a regression model to investigate the impact that information consistency has on sales predictions. Based on our empirical analysis, we found that both image-based and text-based information influenced consumers’ purchase decisions but that the former influenced their purchase decisions about “search goods” more and that the latter influenced their purchase decisions about “experience goods” more. Furthermore, congruence between image-and text-based information was positively associated with purchase decisions, which indicates that information congruence impacts products’ sales performance in the digital economy. In this study, we also demonstrate how to apply advanced deep-learning techniques to measure the congruence between different information types.
AB - In a digital economy, different types of information about products communicate their quality and characteristics to prospective consumers. However, it remains unclear which type of information plays the most important role in individuals’ decision-making processes. In this study, we explore the effect that unstructured data has on and the importance of congruence between textual and visual data in consumers’ purchase decisions. We apply a deep neural network model to rank the importance of different information types and use a regression model to investigate the impact that information consistency has on sales predictions. Based on our empirical analysis, we found that both image-based and text-based information influenced consumers’ purchase decisions but that the former influenced their purchase decisions about “search goods” more and that the latter influenced their purchase decisions about “experience goods” more. Furthermore, congruence between image-and text-based information was positively associated with purchase decisions, which indicates that information congruence impacts products’ sales performance in the digital economy. In this study, we also demonstrate how to apply advanced deep-learning techniques to measure the congruence between different information types.
KW - Deep Learning
KW - Digital Economy
KW - Image
KW - Information Congruence
KW - Neural Network Model
KW - Text
KW - Unstructured Data
UR - http://www.scopus.com/inward/record.url?scp=85096979587&partnerID=8YFLogxK
U2 - 10.17705/1CAIS.04708
DO - 10.17705/1CAIS.04708
M3 - Article
AN - SCOPUS:85096979587
VL - 47
SP - 165
EP - 187
JO - Communications of the Association for Information Systems
JF - Communications of the Association for Information Systems
SN - 1529-3181
M1 - 8
ER -