Abstract:

This study aims to address the growing need for effective quality assessment in e-commerce platforms, driven by the rapid evolution of online transactions globally. The absence of a structured approach to categorize and analyze shopping app reviews based on specific quality dimensions represents a significant gap in the existing literature, emphasizing the importance of this research endeavor. Methodologically, we propose a taxonomy framework designed to classify users' reviews from shopping apps into five quality-centered categories: system quality, information quality, service quality, product quality, and delivery quality. Leveraging an initial dataset of 552 instances, we employ and compare five common multi-label classification models: binary relevance, classifier chain, label powerset, multi-label k nearest neighbors, and multi-label decision tree. Our findings reveal promising outcomes in terms of classification accuracy and performance metrics. Employing binary relevance with logistic regression and BERT embedding resulted in a Hamming loss of 0.177, Macro-F1 score of 0.739, and Micro-F1 score of 0.744, Macro Recall of 0.722, and Micro-Recall of 0.728. Notably, the use of binary relevance with SVM and Glove embedding demonstrated the highest Macro-Precision and Micro-Precision at 0.944 and 0.933, respectively. Concerning the Zero-One Loss metric, the better outcome attained was 0.572 by using the Classifier Chain with Glove embeddings. These findings highlight the efficacy of our proposed taxonomy and classification methodology in effectively evaluating and categorizing shopping app reviews based on quality dimensions, thus contributing significantly to the field of e-commerce analytics and user experience enhancement.