Title :
Research on Event Analysis Approach for the Identification of Aircraft Large Vertical Load
Author :
Zhao Yu ; Chen Rui ; Huang Si-ming ; Xu Bao-guang
Author_Institution :
Inst. of Policy & Manage., Chinese Acad. of Sci., Beijing, China
Abstract :
Aircraft large vertical load means the aircraft is weighted above the normal, which may harm the air safety. So it is crucial to do the research on how to choose suitable models and methods to identify whether the aircraft got a large vertical load before landing and adjust immediately. The MWW nonparametric test is used to extract the distinct features. Then, compare the most commonly used models and choose the k-nearest neighbors and support vector machines with dynamic weight consideration to classify, both of them ensure the accuracy and make sense according to the feature of the dynamic data and comparison of some classify methods. The results indicate that the two methods make no difference on this problem but the model v-SVC in support vector machine seems more adaptable when the problem expands.
Keywords :
aerospace computing; air safety; aircraft; data mining; feature extraction; pattern classification; support vector machines; MWW nonparametric test; air safety; aircraft large vertical load identification; data mining; event analysis approach; feature extraction; k-nearest neighbors; support vector machines; v-SVC model; Accuracy; Aircraft; Atmospheric modeling; Kernel; Support vector machine classification; Training;
Conference_Titel :
Management and Service Science (MASS), 2010 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-5325-2
Electronic_ISBN :
978-1-4244-5326-9
DOI :
10.1109/ICMSS.2010.5575672