基于深度学习的人工智能空气质量预报系统构建

针对现有空气质量预报系统存在预报精度低、人工经验辅助、适用范围单一等问题,利用深度学习方法在分析数据内在特征方面表现出的优异性能,结合多源数据融合技术,设计了基于深度学习的空气质量预报系统实现方案。通过对多源数据集的实时制作更新、分析空气质量演变的时空特性、定义和拟合深度学习模型并部署于服务器等关键技术的研究,最终实现了空气质量的多尺度、高精度实时预报服务和预报结果可视化服务。应用结果表明,基于深度学习的空气质量预报系统具有更高的预报精度和更优良的应用效果,可提高预报效率,为空气质量预报服务提供一种新型、高效的实现方式。

Construction of Artificial Intelligence Air Quality Prediction System Based on Deep Learning

Aiming at the problems (e.g.low prediction accuracy, manual experience assistance, single application scope and etc.) of the existing air quality prediction systems, this paper uses the superior performance of deep learning method in analyzing the internal characteristics of data, combined with the multi-source data fusion technology to design the implementation scheme of air quality prediction system.The key technologies are studied, such as real-time production and update of multi-source data sets, analysis of spatial-temporal characteristics of air quality evolution, definition and fitting of deep learning model and deployment in servers.Based on the research, multi-scale and high-precision real-time forecast service and visualization of forecast results of air quality are realized.The application results show that the proposed air quality prediction system based on deep learning has higher prediction accuracy and better application effect than the traditional methods, which can improve the efficiency of forecasting, and provides a new and efficient way for air quality prediction service.