时间：2024年6月12日（周三）11:30-12:30
地点：东北财经大学劝学楼220会议室
报告人：张连奎
摘要：The issue of air pollution, specifically PM2.5 pollution, is progressively worsening, and accurate long-term and high-precision prediction of PM2.5 concentrations is crucial for effective PM2.5 emission control strategies and public health management. Deep learning methods have demonstrated their effectiveness in air quality prediction in recent years. In this study, we propose a deep learning model (HAE) that incorporates a hybrid attention mechanism and embedding method to enhance the accuracy of PM2.5 concentration prediction. Our model integrates various types of attention mechanisms into both input and output sequences to address the challenge of long-term memory retention, enabling better capture of long-range dependencies. Additionally, an embedding method is employed to process input features, allowing end-to-end training for each feature value and yielding higher-dimensional representations that are more comprehensive and faithful. To validate our proposed model, we conducted a 72-hour multi-step PM2.5 concentration prediction experiment using the Beijing dataset from China. The results show that the prediction accuracy of the model improves from 22.47% to 76.19% compared to traditional statistical learning algorithms and machine learning algorithms. Furthermore, when compared with common attention mechanism algorithms and Transformer models, our model exhibits respective improvements in accuracy by 17.35% and 9.20%.
报告人简介：张连奎，东北财经大学公共管理学院，大连理工大学博士，主要研究方向为大数据分析，环境政策与管理，相关研究成果发表在Journal of Environmental Management、Environmental Research、Journal of Systems Science and Systems等期刊。
