Uncovering Direct Influence Networks of Takeoff Weight: Network Science and Information Theoretic Approach

Accurate estimation of aircraft takeoff weight (TOW) from operational data is critical for performance analysis, fuel modeling, and environmental assessment, yet true mass information is rarely available in open datasets. This paper proposes an information-theoretic, network-science framework to uncover the direct influence structure underlying TOW using ADS-B–derived and operational variables. Mutual information and conditional mutual information are used to construct a dependency network and to suppress indirect associations. The resulting sparse network reveals clear clustering by physical variable type and identifies a limited set of contextual variables directly informative of TOW, highlighting both informative predictors and limitations of current trajectory feature representations.