基于粗糙集神经网络的燃煤发热量预测模型

Forecast of Coal Heat with Rough Sets Theory and Artificial Neural Networks

通过对粗糙集和BP神经网络的分析研究,以专家系统为核心,提出了一种基于粗糙集神经网络的燃煤发热量预测模型;选取影响燃煤发热量的6个参数,利用粗糙集理论对原始信息表进行约简操作,去除冗余的属性和属性值,得到约简规则,并将其作为BP神经网络的输入,对燃煤发热量进行预测;通过分析对比线性回归方法和粗糙集神经网络方法,说明该模型能有效地简化神经网络的网络结构,减少网络的训练步数,提高网络的学习效率,能够较准确地对燃煤发热量进行预测。

Through the research and analysis of rough sets theory and BP neural network, a forecast model based on rough sets theory and BP neural network is put forward which operates on the expert system. To forecast the quantity of coal heat, six parameters affecting the coal heat are selected to build the decision table. And then, rough sets theory is applied to reduce the redundancy of the attribute and attribute value in the decision table. Finally, the reduction results are transformed into rules, which are used as input of the BP neural networks to build the forecasting model. By analyzing and contrasting with the linear regressive method, it is confirmed that the model could not only reduce structure and training step of the neural network effectively, but also could improve the learning efficiency and forecast the quantity of coal heat well.