This study aims to deeply understand customer experiences toward Internet of Things (IoT) applications in retail by developing machine learning models for aspect-based sentiment analysis (SA). It includes creating a related terms dictionary and proposing implications for retail businesses in Vietnam based on these analyses. The ultimate goal is to gain insights into customer opinions and assist administrators in formulating effective digital transformation and business strategies within the Vietnamese market.
Initially, this research uses qualitative methods to identify different aspects of customer experience at stores equipped with IoT applications. Then, quantitative methods were applied through classification machine learning models which were trained on the annotated data set to classify comments into aspects and sentiments. Finally, the classification results were analyzed and visualized to draw implications about customer opinions of these stores.
This study collected 77,042 customers’ comment from potential and actual customers who have ever shopped at retail stores with IoT applications deployed worldwide, identified ten new aspects of customer experience in this field and built a dictionary of related terms. Furthermore, this study contributed two efficient ensemble models with an accuracy of 81% and 89% for analyzing aspects and customer sentiments, respectively. This study also proposes implications for managers regarding the use of IoT technology in retail stores to improve shopping experiences for customers.
This study’s findings help managers develop appropriate digital transformation and business strategies for integrating IoT technology into retail stores, especially for retail businesses in the Vietnamese market based on the analysis results and proposed model.
