Emotion-Driven Design of Rose Chairs: A Female Preference Prediction Model Integrating EMD-KPCA-LSTM Hybrid Architecture

To elucidate the complex interaction between female users \u2019 affective-cognitive preferences and the generative mechanism of rose chair design, this study investigates the modern translation mechanism of traditional rose chair design while achieving predictive perceptual modeling in design outcomes. We introduce a multimodal framework integrating KJ technique, exploratory factor analysis (EFA), K-means clustering algorithm, and triangular fuzzy number theory, establishing a dynamic coupling mechanism between users\u2019 visual perception hierarchy and multidimensional preference construction. Eye- tracking technology was employed to capture users \u2019 visual trajectory data, which were transformed into a preference perception repository through a dual-coding strategy that combines semantic differential scales with eye-tracking metrics. This approach deciphers the latent relationships between three core aesthetic dimensions (elegance, delicacy, comfort) and their corresponding morphological mechanism parameters.\r\nThe results demonstrate that our proposed hybrid model \u2014 integrating empirical mode decomposition (EMD), kernel principal component analysis (KPCA), and long short-term memory (LSTM) networks (EMD-KPCA-LSTM)\u2014outperforms conventional methods such as BP neural networks, support vector machines (SVMs), and convolutional neural networks (CNNs). Our model exhibits superior predictive accuracy, enhanced generalization capability through cross-validation, and improved robustness against noisy perceptual data. By constructing a quantifiable bridge between affective-cognitive responses and parametric form generation, this study advances user-centered design theory and offers transformative potential for intelligent design systems to balance cultural semantics and user-centric aesthetics in the innovation of traditional furniture.