New York: The last-minute information about delays of your scheduled flight may soon become a thing of the past, thanks to a new computer model that could lead to more accurate and quicker flight delay predictions.
“Our proposed method is better suited to analyse datasets with categorical variables (qualitative variables such as weather or security risks instead of numerical ones) related to flight delays,” said the lead author of the study, Sina Khanmohammadi from Binghamton University – State University of New York.
“We have shown that it can outperform traditional networks in terms of accuracy and training time (speed),” Khanmohammadi said.
While the research would not eliminate delays, the researchers believe it will help airlines inform travellers quicker and more accurately about problems.
“Airlines can use the proposed method to provide more accurate delay information to the customers, and hence gain customer loyalty.”
The new model could also help smaller regional airports become more efficient and handle more flights per day.
“Air traffic controllers at a busy airport can also use this information as a supplement to improve the management the airport traffic,” Khanmohammadi said.
Currently, flight delays are predicted by artificial neural network (ANN) computer models that are backfilled with delay data from previous flights.
The research team introduced a new multilevel input layer ANN to handle categorical variables with a simple structure to help airlines easily see the relationships between input variables (such as weather) and outputs (flight delays).
Researchers trained the new model to pick up on 14 different variables – including day of the week, origin airport, weather and security – that affected arrival times for 1,099 flights from 53 different airports to John F. Kennedy airport in New York City.
The new system then predicted delays for hypothetical flights projected to arrive at JFK at 6.30 p.m. on January 21 from a variety of origins and under a variety of conditions.
The new model, described in the journal Procedia Computer Science, predicted the length of delays with about 20% more accuracy than traditional models and required about 40% less time to come to those conclusions.