摘要
针对底板破坏带的精度问题提出新的预计模型,通过搜集众多矿井的实测数据,应用多元统计分析算法,在支持向量机的基础上建立预计模型。采用果蝇优化算法对预计模型进行优化训练,建立FOA—SVM预计模型。利用实测数据对模型的预计结果进行检验,预计结果准确,比遗传算法模型、粒子群模型的预计结果稳定性更好和精度更高。
A new model was raised for the accuracy of the floor with destruction.Collecting experimental data of numerous mines and applying multivariate statistical analysis algorithms,the model was built on the basis of the supporting vector ma-chines.The FOA-SVMprediction model was established using the Drosophila optimization algorithm.The prediction results was tested by the measurement data,with accurate prediction results.The new model was better than the genetic algorithm model and the particle swarm model,with better stability and higher accuracy.
出处
《世界科技研究与发展》
CSCD
2014年第5期494-497,共4页
World Sci-Tech R&D
基金
国家自然科学基金(51174109)资助
关键词
底板破坏带
支持向量机
果蝇优化算法
模型优化
多元统计分析
destruction line of floor
support vector machine
drrosophila
model optimization
multivariate statistical analysis