A three-dimensional integrated evaluation system for product competitiveness under complex demands: An LDA and PSO-FAHP based hybrid optimization approach

Abstract

Given the lack of interpretability and limited adaptability of traditional product evaluation methods due to their reliance on expert experience, this study proposes an integrated evaluation system that integrates LDA-PSO-FAHP. LDA topic modeling analyzes user-generated content, extracts potential product attributes and consumer needs, and constructs a three-dimensional evaluation system. Secondly, a PSO-FAHP hybrid decision model is developed, and the initial weights are generated using fuzzy hierarchical analysis combined with particle swarm optimization to correct the judgment matrix dynamically. Then, the design case of elderly walkers is used as an empirical object, and verification shows that the system evaluation results are significantly positively correlated with the sales growth rate of the B2B platform and the user praise rate. Finally, comparative experiments prove that this method is superior to traditional methods in terms of consistency of judgment matrix and accuracy of weight calculation. The system aims to provide a decision-making tool for the manufacturing industry that considers both market semantic analysis and design parameter optimization and helps product development transform from a subjective experience-driven to a data-model collaborative decision-making paradigm.