Cooperative training-based city region air quality estimation method

The invention relates to a cooperative training-based city region air quality estimation method. The method makes full of spatial features of a region, such as features of the traffic condition, the road network structure and the like of the region, modeling is performed on feature vectors of the region based on a semi-supervised learning method of cooperative training of multiple classifiers, and multiple classifiers are obtained through learning; then pruning is performed on the classifiers, thereby forming a final combined classifier; and a pruned combined classifier model is used to perform air quality grade estimation on the region having no air monitoring station. The method can estimate the air quality grade according to various differences in space between a region with an air monitoring station and a region without an air monitoring station under a circumstance that air monitoring stations are limited, and an estimation result is accurate.