One effective method of outlier detection in flight data

Flight data is a kind of time correlated observations to each other as a time series in nature, which is not comply with some certain models, detecting outliers from the time series is a big challenge. In this article an effective two-sided median filtering method to detect outliers in flight data is proposed, which makes use of the median computed from a local data point´s certain neighborhood window and the reasonable threshold to compare with the difference between the median and the observed data value. Then outliers are replaced by the appropriate medians in order to prevent from missing data in time series. It is essential of putting forward a credible and universal evaluation method to evaluate the result of outlier detection in flight data. This outlier detection method is efficiently and effectively used for preprocessing flight data and detecting the outliers in flight data.