A BP Neural Network Algorithm for Multimedia Data Monitoring of Air Particulate Matter

Table 2

Description of model construction.

Model input layer	4 nodes PM2.5PM10 humidity and temperature
Model hidden layer	The maximum number of layers is 2, and the maximum number of nodes in each layer is 10
Model output layer	1 node PM2.5
Model bias quantity	1
Model training data	The hourly average data of air monitoring equipment and state control points are randomly selected at a ratio of 0.9
Model test data	Hourly average data of air monitoring equipment and state control points with the remaining ratio of 0.1
Test evaluation function	According to the MSE evaluation standard of mean square error function, the optimal model of hidden layer is selected
Validation data	Hourly average data of air monitoring equipment and state control points