Research on New Energy Vehicle Sales Model Based on Attention-Seq2Seq Architecture

With worsening energy and environmental issues, new energy vehicles (NEVs) have emerged as the automotive industry’s future that aims to address high energy use and carbon emissions of traditional fuel vehicles. However, NEV research mostly focuses on theoretical analysis due to short industry history, limited data, and incomplete systems, hindering accurate sales prediction. Online reviews now offer a new perspective for forecasting by influencing consumer decisions. Based on consumer behavior and neural network theories, this study reviews relevant literature, selects NEV sales-influencing factors (economy, technology, policy, consumers, with preprocessed crawled online reviews), constructs an index system screened via grey correlation analysis, and establishes four models (GRU, Seq2Seq, Attention-GRU, Attention-Seq2Seq) for training and testing. Results show online reviews, battery output, and public charging piles effectively support NEV sales prediction. The Attention-Seq2Seq model outperforms the other three across all metrics.