摘要
利用模糊贝叶斯网络模型和小波时间序列模型进行高炉铁水含硅量预报,预测命中率(预测值与实际值误差±0.1)分别为84%和76%,基本反映高炉生产实际.对比分析两个模型的预报结果以及高炉内部反应机理的特点,结合两个模型优势,提出组合预报模型,把预报命中率提高到92%,具有生产应用价值.
The hit rates （ the error between the predictive value and the actual value is ± 0. 1 ） of using fuzzy bayesian network and wavelet time series model to predict the silicon content in BF molten iron are respectively 84% and 76%, which basically reflect the BF actual production. After detailed comparison and analysis of the two prediction results, with internal BF characteristics of the reaction mechanism, a combined forecasting model is proposed. And the method can increased the hit rate to 92% and therefore is applicable to the real production.
出处
《嘉兴学院学报》
2008年第6期11-14,共4页
Journal of Jiaxing University
关键词
铁水含硅量
模糊贝叶斯网络
小渡时间序列
预报
silicon content in molten iron
fuzzy bayesian network
wavelet time series
prediction
