How to prioritize perceived quality attributes from consumers' perspective? Analysis through social media data

Social media data is gaining attention as consumers become accustomed to sharing and finding product perceptions on social media. Perceived quality is consumers' subjective perceptions, and it is important for manufacturers to prioritize perceived quality attributes. However, existing studies mainly use survey data, which is prone to bias, and lack analysis of why perceived quality arises. We propose a three-stage framework to prioritize perceived quality attributes using social media data based on text mining techniques. First, a deep-learning approach is used to identify perceived quality; second, the group perceived quality, attribute importance, and quality category of the attribute are synthetically analyzed to quantify the perceived quality, and perception causes are mined; finally, importance-performance analysis is used to prioritize attributes and a bottom-up cause chart is built. In the case study, an automobile dataset is crawled to apply the proposed framework and the results are validated in a user experiment.