The haze level forecast based on the hybrid neural networks model

Haze is a weather phenomenon, which is very harmful to the environmental protection or the public health. Therefore, it is necessary to forecast upcoming haze weather so that the protective measures can be taken to minimize the haze´s negative impact. A hybrid neural networks model is proposed in this paper for the forecast of the haze level, which consists of two individual neural networks - one is a standard neural network and another is a risk neural network. Experimental results show that the hybrid model is superior to both the standard neural network and the risk one, which provide a support for our work.