Identifying Purchasing Factors in Online Flea Markets Considering Thumbnail Images

Authors

  • Emi Iwanade Graduate School of Information and Telecommunication Engineering, Tokai University, Japan
  • Yoshihisa Shinozawa Faculty of Science Technology, Keio University, Japan
  • Kohei Otake Faculty of Economics, Sophia University, Japan

DOI:

https://doi.org/10.47852/bonviewJDSIS52024073

Keywords:

online flea markets, images, machine learning, gradient boosting, consumer behavior

Abstract

In recent years, the market for online flea markets, which are consumer-to-consumer (C2C) services where goods are bought and sold among users, has been expanding. In such services, sellers (individuals that offer goods or services for sale) create product details, including price, condition, shipping method, and images, when listing their items for sale. We consider the product image to be the first thing users see when selecting a product as a thumbnail, significantly impacting their purchase decisions. In this study, we proposed a discriminant model of purchase decisions for online flea market data to clarify the factors influencing these decisions based on product details. Specifically, we used metadata such as price and delivery method, along with image labels for product thumbnails, as features. We created and compared models with three patterns: metadata only, image labels only, and a combination of metadata and image labels. We created four types of models in this study: logistic regression, decision tree, gradient boosting, and random forest. We selected the models based on their accuracy evaluations. Our analysis revealed that the model using both metadata and image labels as features, combined with the gradient boosting method, had the highest accuracy. The partial dependence plots of the selected models highlighted the features important for users' purchase decisions.

 

Received: 10 August 2024 | Revised: 8 February 2025 | Accepted: 14 March 2025

 

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

 

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

 

Author Contribution Statement

Emi Iwanade: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Writing – original draft, Writing – review & editing, Visualization. Yoshihisa Shinozawa: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Writing–original draft, Writing – review & editing, Visualization. Kohei Otake: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition.


Downloads

Published

2025-03-26

Issue

Section

Research Articles

How to Cite

Iwanade, E., Shinozawa, Y., & Otake, K. . (2025). Identifying Purchasing Factors in Online Flea Markets Considering Thumbnail Images. Journal of Data Science and Intelligent Systems. https://doi.org/10.47852/bonviewJDSIS52024073

Funding data