【摘要】：隨著民用飛機和航線數(shù)量的不斷攀升,民航業(yè)進入了大數(shù)據(jù)時代,如何有效地檢測異常數(shù)據(jù)并且避免它們對分析結果的不利影響是研究的焦點。本文圍繞四分位法和目前民航界慣用的標準差法展開討論,主要從數(shù)據(jù)集分布、異常值比例、異常值檢測能力、抗擾動性和可視化效果5方面對四分位法和標準差法進行分析。最后通過快速存取記錄器(Quick access recorder,QAR)數(shù)據(jù)分析論證,四分位法在上述5方面都表現(xiàn)出比標準差法更好的性質,四分位法更適合應用在民航告警領域。
[Abstract]:With the increasing number of civil aircraft and airlines, the civil aviation industry has entered the era of big data, how to effectively detect abnormal data and avoid their adverse effects on the analysis results is the focus of research. In this paper, the quartile method and the standard deviation method used in civil aviation are discussed. The quartile method and standard deviation method are analyzed from five aspects: data set distribution, outlier ratio, outlier detection ability, anti-disturbance and visualization effect. Finally, through the data analysis of (Quick access recorder,QAR, the quartile method is better than the standard deviation method in the above five aspects, and the quartile method is more suitable for civil aviation alarm.
【作者單位】： 中國民航科學技術研究院航空安全研究所;
【基金】：國家自然科學基金民航聯(lián)合基金(U1533120)資助項目;國家自然科學基金(11371365,11301535)資助項目
【分類號】：V328
，

本文編號：2325392