The Digital Reconstruction of Traditional Chinese Opera: Interactive Innovative Design of Quanzhou String Puppetry

To enhance the ability of task issuers in crowdsourced design to accurately identify and describe design requirements, this study analyzes a large number of online reviews to extract market demands and utilizes large language models to generate more accurate and comprehensive design requirements. Online review texts from e-commerce platforms were collected, and the evaluation subjects within the reviews were identified and analyzed. These subjects were then categorized using spectral clustering, user satisfaction was calculated, and design requirements were generated through dialogue with large language models. This method systematically analyzes user feedback from online reviews, identifies and clusters evaluation subjects, and further understands and generates user requirements. Experiments demonstrate that this method effectively generates crowdsourced design requirements. This research effectively integrates the extraction of design requirements from online review texts with the recognition and generation of these requirements using large models. It achieves precise localization and efficient generation of design needs, providing a new implementation approach for crowdsourced design.