基于因子分析的航班撤轮挡时刻预测方法

Flight off-block time prediction based on factor analysis

为提升机场、航空公司、空管等部门的协同决策效率,提出构建基于因子分析和多元线性回归的航班撤轮挡时刻预测模型。通过因子分析确定与航班撤轮挡时刻相关的里程碑事件,将其作为自变量,建立多元线性回归模型预测航班撤轮挡时刻。实验结果表明,在误差范围±5min和±10min内,该模型的平均预测准确率分别可达70%和90%以上,高于A-CDM经验方法和基于主成分分析的航班撤轮挡时刻预测算法。

Flight off-block time prediction can help airports and airlines as well as air traffic control make efficient collaborative decision.To this end,an off-block time prediction model was proposed based on factor analysis and multiple linear regressions.Factor analysis was used to identify milestone events related to off-block time.A multiple linear regression model was established based on those milestone events.Experimental results show that within the error ranges of±5 and ±10 minutes,the average prediction accuracy rates raise up to 70% and 90% respectively.Compared with the A-CDM empirical estimation method and the combination of principal component analysis and multiple linear regression,the proposed algorithm performs better in terms of prediction accuracy.