Abstract: A neural network including a set of input nodes may consume a respective stream of time-series data recorded during a flight of a flying aircraft, each stream of time-series data representing measurements of a respective flight parameter captured by a sensor at various time-steps of the flight. A training circuit set may train the neural network to predict a future measurement of the flight parameter. Training the neural network may include comparing a predictive value from the neural network to a measured value of a flight parameter and modifying structural components of the neural network to bring the predictive value closer to the measured value. A parameter acquisition circuit set may acquire time-series data of a flight parameter. A prediction circuit set may apply the time-series data to the trained neural network to predict the next measurement for the flight parameter in the time-series data.

Abstract: Various embodiments for analyzing flight data using predictive models are described herein. In various embodiments, a quadratic least squares model is applied to a matrix of time-series flight parameter data for a flight, thereby deriving a mathematical signature for each flight parameter of each flight in a set of data including a plurality of sensor readings corresponding to time-series flight parameters of a plurality of flights. The derived mathematical signatures are aggregated into a dataset. A similarity between each pair of flights within the plurality of flights is measured by calculating a distance metric between the mathematical signatures of each pair of flights within the dataset, and the measured similarities are combined with the dataset. A machine-learning algorithm is applied to the dataset, thereby identifying, without predefined thresholds, clusters of outliers within the dataset by using a unified distance matrix.

Abstract: A neural network including a set of input nodes may consume a respective stream of time-series data recorded during a flight of a flying aircraft, each stream of time-series data representing measurements of a respective flight parameter captured by a sensor at various time-steps of the flight. A training circuit set may train the neural network to predict a future measurement of the flight parameter. Training the neural network may include comparing a predictive value from the neural network to a measured value of a flight parameter and modifying structural components of the neural network to bring the predictive value closer to the measured value. A parameter acquisition circuit set may acquire time-series data of a flight parameter. A prediction circuit set may apply the time-series data to the trained neural network to predict the next measurement for the flight parameter in the time-series data.

Abstract: Various embodiments for analyzing flight data using predictive models are described herein. In various embodiments, a quadratic least squares model is applied to a matrix of time-series flight parameter data for a flight, thereby deriving a mathematical signature for each flight parameter of each flight in a set of data including a plurality of sensor readings corresponding to time-series flight parameters of a plurality of flights. The derived mathematical signatures are aggregated into a dataset. A similarity between each pair of flights within the plurality of flights is measured by calculating a distance metric between the mathematical signatures of each pair of flights within the dataset, and the measured similarities are combined with the dataset. A machine-learning algorithm is applied to the dataset, thereby identifying, without predefined thresholds, clusters of outliers within the dataset by using a unified distance matrix.