Forecasting Clinker Strength Based on Rough Set and Neural Network

Proposed cement clinker strength forecasting model based on rough set & neural network by analyzing the dangers of clinker strength detecting lag and the current forecasting means & methods with many shortcomings and limitations. The model makes good use of the advantage of rough set identifiable matrix attribute reduction and neural network good at dealing with non-linear problem. After training network by samples from cement factory, 3rd and 28th day forecasting clinker strength error is between 0-5 percent. The result illustrates that predicting by rough set & neural network model is reliable. It can be applied to guide further cement production and ensure the quality of cement.