Air pollution is worsening almost everywhere in the world. According to the Health Effects Institute (HEI), more than 95% of the world population breathe polluted air, toxic to their cardiovascular and respiratory health, which caused the death of 4.2 million people worldwide in 2016. As a result, the air pollution has become one of the leading causes of death worldwide. Therefore, an early cost-efficient warning system based on precise forecasting tools must be put in place to measure and avoid the harmful effects of exposure to the main air pollutants. Thus, it is essential to obtain reliable analytical information on air quality in a specific time and place. This paper focuses on monitoring air quality using a distributed intelligence which is a cost-efficient solution that enables a flexible prediction process distributed within a network of nodes and devices using a cross-platform solution. The suggested architecture enables collaborative learning along with collective knowledge graph building and knowledge sharing using the state of the art in Internet of Things, distributed machine learning, and ontologies. The proposed architecture suggests a flexible prediction system personalized for each node based on its need of information. Similar nodes get together for collective learning which allows for resource optimization, knowledge reusability, and device interoperability. The paper describes the modeling framework of distributed intelligence monitoring and analysis system designed for urban regions.

空气污染几乎在世界各地都在恶化。根据健康影响研究所（HEI）的数据，全球超过95％的人口呼吸被污染的空气，这对他们的心血管和呼吸系统健康有毒，导致2016年全球420万人死亡。结果，空气污染成为全球死亡的主要原因之一。因此，必须建立一个基于精确预测工具的具有成本效益的早期预警系统，以测量和避免暴露于主要空气污染物的有害影响。因此，至关重要的是在特定的时间和地点获得有关空气质量的可靠分析信息。本文着重于使用分布式智能监控空气质量，这是一种经济高效的解决方案，可使用跨平台解决方案在节点和设备的网络内实现灵活的预测过程。所建议的体系结构可以使用物联网，分布式机器学习和本体中的最新技术实现协作学习以及集体知识图的建立和知识共享。所提出的体系结构提出了一种基于每个节点的信息需求而个性化的灵活预测系统。相似的节点聚集在一起进行集体学习，从而实现资源优化，知识可重用性和设备互操作性。本文描述了为城市地区设计的分布式智能监视和分析系统的建模框架。