Product design opportunity identification through mining the critical minority of customer online reviews

Online reviews that contain customer requirements and expectations are valuable for product design opportunity identification (PDOI) in customer-centric design. As most reviews are non-informative and repetitive, the existing approaches may not be effective for PDOI. Besides, online reviews containing design opportunities tend to be the critical minority, which can be characterized as outliers in a review dataset (RDT). Motivated by this observation, this study develops an online review mining approach based on outlier detection technique to identify the critical minority of online reviews for PDOI. First, unstructured online reviews are modelled as a RDT from a product improvement perspective, and the metadata and information attributes of review are involved. Then, a non-parametric weighted neighborhood information network (WNIN)-based outlier detection method is investigated to determine outlier reviews. Finally, a real case study of the PDOI for Mi 10 is implemented to elaborate the feasibility and effectiveness of the proposed methodology.