Study on Early Warning for Coal Industry Security Based on BP Neural Network and Rough Sets

Aiming at the shortages of the existing data-mining model for forecasting the industry security, a classification model based on rough sets and BP neural network (BPNN) is put forward in this paper. First, the theory of rough set is applied to pick up and reduce the index attributes. Then, the training samples are sent to the BPNN to train and learn. After that, the sorts of the coal industry security in test samples are differentiated. The test results indicate that the classification model based on rough sets and BPNN shows higher forecast precision than the traditional ones and it is more efficient and practical. The result of forecasting shows China´s coal industry in 2015 is the basic security status.