Construction of a Dynamic Hybrid Model for Analyzing the Demand of Chinese NEV Consumers: Based on DS-Kano and C2SLM-Kano

Against the backdrop of the rapidly expanding global new energy vehicle (NEV) market in 2024–2025, consumer demand behavior exhibits dynamic regional variability, challenging traditionally static and homogeneous analytical approaches. This research introduces a dynamic hybrid model that combines dynamic-scenario Kano (DS-Kano) and correlation clustering sparse learning Kano (C2SLM-Kano) to tackle this issue. DS-Kano combines long short-term memory (LSTM) and deep reinforcement learning (DRL) for minute-level responsiveness to demand shifts. C2SLM-Kano uses graph convolutional network (GCN) clustering to quantify regional and cultural differences. Empirical validation showed that DS-Kano achieves 7.3 min latency (vs. 45 min for BERT-TCBAD-Kano and 360 min for traditional Kano), C2SLM-Kano reaches 89.2% cross-regional prediction accuracy, and the hybrid model improves demand identification F1-score by 30%, shortening product iteration time by 67% as shown by simulation-based evidence. Robustness checks confirm stability under data perturbations, algorithm swapping, and cross-regional validation. Practical implications should be considered by manufacturers to guide adaptive R&D and by policymakers to design targeted subsidies with case-informed estimates of policy effectiveness.