油耗预测中显著影响参数提取方法的仿真

摘要

在油耗预测中显著影响参数提取问题的研究中,QAR记录了大量与飞行油耗相关的真实过程参量.由于参数之间呈非线性关系,若将相关参数全部输入到模型中,将导致油耗预测模型结果复杂,运行时间长,预测精度低.针对以上不足,提出了一种粗糙集属性约简的油耗估计显著影响参数的提取方法,首先是对QAR数据进行预处理,然后进行离散化,离散化采用了信息熵连续属性的离散化算法,进一步提高了数据分析的可靠性,再通过Rough set软件得出核心属性,经实验,新方法在提取油耗显著影响因素中有效,而且在处理QAR数据中有着广阔的应用前景.%QAR recorded a large number of parameters of the flight fuel consumption process of real operation,and the relations between them are nonlinear.If all parameters which are relevant to fuel consumption are input to estimation model,it must be time-consuming,meanwhile the result would be complex and not accurate.For the above shortcomings,in this paper,an extraction method was presented based on rough set attribute reduction for significant impacting parameters of fuel consumption estimation.Firstly,QAR data were pretreated,then discredited using continuous attribute discrimination algorithm based on information entropy in order to improve the reliability of the results of data analysis,and finally,imported to the rough set software to draw the core attributes.The experiment proves that the method is effective for the extraction of significant fuel consumption afecting factors,and has broad application prospects in processing QAR data.