Graphical abstract Display Omitted
Highlights A method for finding similar days in air traffic flow management initiative planning is proposed. Features describing conditions on calendar days are defined. The importance of these features is determined by modeling initiative implementation. Cluster analysis reveals minimal structure in feature data.
Abstract This article describes a methodology for selecting days that are comparable in terms of the conditions faced during air traffic flow management initiative planning. This methodology includes the use of specific data sources, specific features of calendar days defined using these data sources, and the application of a specific form of classification and then cluster analysis. The application of this methodology will produce results that enable historical analysis of the use of initiatives and evaluation of the relative success of different courses of action. Several challenges are overcome here including the need to identify the appropriate machine learning algorithms to apply, to quantify the differences between calendar days, to select features describing days, to obtain appropriate raw data, and to evaluate results in a meaningful way. These challenges are overcome via a review of relevant literature, the identification and trial of several useful models and data sets, and careful application of methods. For example, the cluster analysis that ultimately selects sets of similar days uses a distance metric based on variable importance measures from a separate classification model of observed initiatives. The methodology defined here is applied to the New York area, although it could be applied by other researchers to other areas.
A methodology for identifying similar days in air traffic flow management initiative planning
2016-05-22
15 pages
Article (Journal)
Electronic Resource
English
A methodology for identifying similar days in air traffic flow management initiative planning
Online Contents | 2016
|Identifying similar days for air traffic management
Elsevier | 2017
|Short-term traffic flow rate forecasting based on identifying similar traffic patterns
Online Contents | 2016
|Short-term traffic flow rate forecasting based on identifying similar traffic patterns
Online Contents | 2015
|