A model of discovering customer opinions about IoT applications for retail stores based on sentiment analysis and ensemble learning method

Abstract

IoT (Internet of Things) has revolutionized the retail industry by providing an optimal shopping experience and improving business efficiency for enterprises. This technology is widely used at many retail stores worldwide and opens up new opportunities for retail in Vietnam. The research conducts sentiment analysis of customers who have experienced shopping or may be interested in purchasing products in the future from 06 retail stores with IoT applications to identify sentiments including positive, negative, and neutral. The dataset has 77,043 comments collected on websites and social media platforms. The cleaned data will then be used to experiment with 05 machine learning algorithms, including K-Nearest Neighbors (KNN), Decision Tree, Random Forest, Logistic Regression, and Ensemble model. In conclusion, the ensemble model has the highest average accuracy score of 89%. The model and research result provide a valuable reference to help administrators develop appropriate business and digital transformation strategies for applying IoT technology for retail stores, especially in Vietnam.