Author(s)

Moxuan Wang

Corresponding Author

Moxuan Wang

Abstract

This article focuses on AI (Artificial intelligence) in the field of environmental pollution traceability and dynamic risk prediction. At present, the situation of environmental pollution is severe, and traditional methods can not meet the needs of accurate traceability and prediction, so AI technology is introduced. By collecting the data of pollutant concentration, meteorological and geographical information of several monitoring stations in a typical industrial area for three consecutive years, an algorithm model of environmental pollution traceability based on CNN (Convolutional neural network) and a dynamic risk prediction algorithm model based on LSTM (Long-term and short-term memory network) are constructed. The experimental results show that, in terms of environmental pollution traceability model, the model has a traceability accuracy rate of 85%, a recall rate of 82% and a F1 value of 83.4%, and an accuracy rate of 82%, a recall rate of 80% and a F1 value of 80.9% for organic pollutants. It is not difficult to find that the model not only performs well in environmental pollution traceability and dynamic risk prediction, but also brings help in environmental protection decision-making. However, in the face of sudden environmental incidents, the model performance still has room for improvement.

Keywords

Artificial intelligence; Traceability of environmental pollution; Dynamic risk prediction