본문 바로가기
  • Home

Consumer Response Prediction on an Online Gift Platform Using Generative AI-Based Image Features and Machine Learning

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2026, 31(5), pp.263~273
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : March 3, 2026
  • Accepted : May 4, 2026
  • Published : May 29, 2026

Eunji Kim 1 Hanjun Lee 1

1명지대학교

Accredited

ABSTRACT

This study develops machine learning models to predict review count and wish count on the KakaoTalk Gift platform and analyzes the contribution patterns of consumer response variables. Product data from the dessert and kitchen categories were collected, and features were constructed from price, discount rate, satisfaction score, product title-derived variables, and generative AI-based image variables. Only products with at least 10 reviews were used for the analysis. Random Forest, XGBoost, and LightGBM models were compared, and SHAP analysis was used to interpret the direction of feature contributions. The results show that the importance and contribution directions of variables differ by target signal and category, and that image-based variables provide supplementary predictive information for some categories and targets. This study presents a methodological framework for predicting and interpreting consumer responses in an online gifting context by integrating generative AI-based image variables with product data.

Citation status

* References for papers published after 2024 are currently being built.