Using novelty detection methods to identify abnormalities in aircraft flight data

A two phase novelty detection approach to locating abnormalities in aircraft flight data is presented, providing a framework with which to locate the heights at which they occur and quantitatively assess their impact on the flight. The 1st phase identifies abnormalities at certain heights in a flight and the second phase compares all flights to identify the most abnormal. Results are validated against an existing event based method.