We developed a solution to calculate flight lids dinamically based on passenger probability rather than the flight performance using machine learning algorithms and PNR boarding data.
Optimize profit by increasing inventory and reducing deny boarding cost (and satisfaction).
We look at PNR data and not only flight data to determine probability of show/no-show. These are some of the historical data we look at in order to determine the probability of show/no-show and the total flight potential overbooking
Look at some screenshots of the actual app.