论文部分内容阅读
针对复杂的烧结块成分预测问题,提出一种基于过程神经网络和改进灰色系统的铅锌烧结块成分智能集成预测模型.首先利用过程神经网络可充分表达时间序列中时间累积效应、灰色系统可弱化数据序列波动性的特点,分别对烧结块成分进行预测,然后从信息论的观点出发,提出一种确定各预测模型加权系数的熵值递推算法,通过对两个预测模型的预测结果进行加权集成,获得更加准确的铅锌烧结块成分预测结果.结果表明,智能集成模型的预测精度高于单一预测模型,能有效地对烧结块成分进行预测,满足了配料计算对预测精度和数据完备性的要求.
In order to solve the problem of complex sintering block composition, a new integrated forecasting model of lead-zinc sinter cake based on process neural network and improved gray system is proposed.First, the process neural network can be used to fully express the time accumulation effect in time series, and the gray system can be weakened Data sequence of the characteristics of volatility, respectively, to predict the composition of the sintered block, and then from the point of view of information theory to propose a predictive model to determine the weighting coefficient entropy recursive algorithm, through the prediction of the two predictive models for weighted integration The results show that the prediction accuracy of the intelligent integrated model is higher than that of the single prediction model and the composition of the agglomerate can be effectively predicted to meet the requirements of predicting accuracy and data completeness Claim.