Feature Extraction and Classification Accuracy Validation for New Energy Vehicle Target Customer Segments Based on K-means Clustering

To accurately meet the increasingly diversified consumer demands in the new energy vehicle market, this study takes 218 valid survey samples of new energy vehicle consumers collected nationwide as the data foundation, focuses on the core task of "target customer group feature extraction and classification accuracy verification", and jointly adopts the K-means clustering model and the dual-stream SVM-iTransform-GRU-DT integrated regression model for empirical analysis. The research first uses the K-means clustering algorithm to determine the optimal number of clusters as K=2, and accordingly divides consumers into two core customer groups — "technology-sensitive" (accounting for 62%) and "cost-oriented" (accounting for 38%), and systematically sorts out the significant differentiated characteristics of the two groups in terms of socio-economic attributes, core car purchase concerns and consumer behavior preferences; subsequently, the dual-stream SVM-iTransform-GRU-DT integrated regression model is used for quantitative verification of customer group differences. The model’s goodness of fit R² reaches 0.91835, which not only confirms that residence and occupation are the key influencing factors leading to the differentiation of the two customer groups, but also verifies the reliability of the K-means clustering results. The research conclusions can provide practical references for new energy vehicle enterprises to carry out precise marketing, optimize product R&D directions, and for policymakers to targeted improve relevant supportive policies.