Study On ERA Decision Supporting System Based On Data Mining

Environmental Risk Assessment (ERA) is an important aspect of Environmental Impact Assessment (EIA). The aim of environmental risk assessment is to make decision of risk management, that is to say, decision-making is the last step of environmental risk assessment. However, due to the complexity of environmental risk, it is hard for environmental department to make prompt and proper decision of environmental risk management.More and more cities have created a series of environmental monitoring databases based on Internet and a great deal data about environmental quality are accumulated into these databases. Nowadays, the data processing is limited in real-time monitoring, daily report, monthly report, annually report etc, on which ERA is mainly based, nevertheless, the latent value of these data has not been unearthed.In order to solve the problems mentioned above effectively, an ERA decision supporting system based on data mining is provided in this dissertation, the main task of which includes the following three aspects.1. An idea of introducing data mining into ERA is brought forward, then, the model and framework of an ERA decision supporting system are designed. Furthermore, the database and model base of the decision supporting system are developed respectively by using MySQL and Visual Basic, Visual C++.2. Forecasting models of BP neural network and its improved algorithm, Markov model, and trend structure sequence model are designed and all modules of them are implemented by virtue of Visual Basic and Visual C++. For the sake of adapting the modules of BP neural network and its improved algorithm to various data sources, the modules which are respectivelly based on database and data files have been developped.3. In order to verify the validity and feasibility of the forecasting models for the decision supporting system, recorded data of air quality monitoring from Jinan City, recorded data of TSP and SO₂ concentration monitoring from Dalian Xiao Ye Tian Co., Ltd., and recorded data of TOC concentration monitoring from Dong River in Guangdong Province are used for case analysis. The study indicates that results of forecasting, which are on the whole consistent with practical situation, can reflect the changing trend of the specific environmental quality by and large.In a word, it is basically feasible to apply the ERA decision supporting system based on data mining to decision-making of environmental risk management.