Development of a data-driven smart product service system framework utilizing unsurprised learning model

Many studies have addressed traditional product-service system (PSS) design, but a combination of data-oriented PSS with emerging technologies to achieve a Smart PSS that can respond to a continuously changing environment remains absent. Therefore, this study proposes a systematic framework that utilizes text analytic techniques to capture PSS via a data-oriented service blueprint for use in identifying improvement opportunities and proposes an improvement plan merging a PSS design process and Bidirectional Encoder Representations from Transformers (BERT), which can handle context-sensitive services with smart and connected products in a dynamic environment. By utilizing a data-driven service blueprint and unsurprised learning model, a Smart PSS is transformed. Experiment shows this tourism recommendation generates enhanced service quality and customer satisfaction.