Multi-feature fusion-based consumer perceived risk prediction and its interpretability study
2025

Predicting Consumer Perceived Risk in Online Shopping

Sample size: 262752 publication Evidence: high

Author Information

Author(s): Qi Lin, Xie Yunjie, Zhang Qianqian, Zhang Jian, Ma Yanhong

Primary Institution: School of Economics & Management, Beijing Information Science & Technology University, Beijing, China

Hypothesis

This study aims to predict perceived risk in different contexts by analyzing review content and website information.

Conclusion

The developed predictive model for perceived risk achieved high accuracy and identified key features affecting perceived risk in online shopping.

Supporting Evidence

  • The model achieved a precision of 84%, recall of 86%, and F1 score of 85%.
  • Quality, functionality, and price were identified as key features affecting perceived risk for electronic products.
  • Significant differences in feature characteristics were found between high-risk samples and normal samples.

Takeaway

This study helps understand what makes people worried when shopping online, like product quality and price.

Methodology

The study used a dataset of online reviews and employed machine learning techniques, including KeyBERT-TextCNN for feature extraction and PCA-K-medoids-XGBoost for risk prediction.

Limitations

The model has limitations related to data representativeness and the comprehensiveness of feature selection.

Digital Object Identifier (DOI)

10.1371/journal.pone.0316277

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication