- Job Hunting - February 2, 2017
- Don’t worry: there is life after graduating! - November 30, 2016
- I have a dream and three tips to make it happen - June 14, 2016
- We Are Free to Make Our Own Decisions! Really? - April 5, 2016
- Dating and Mating through the Eyes of an Economist - March 2, 2016
- Career prospect: just graduated Tessa van Acht - January 27, 2016
- Keep in Touch without a Mobile Network - December 8, 2015
A serious price increase of airfare tickets is shown during the holidays in December, says Alexandra Talty in Forbes magazine. This is not surprising, but paying 40% more for a ticket for flying to your family for Christmas than when you would go two weeks earlier is quite a lot. The reasoning behind the higher prices seems logical to us: people have to be at a certain place at a certain date and time and therefore are willing to pay more.
Most of us try to avoid these price increases by booking very early, long before the day of travel. But there are more factors that determine airfare prices. As the World Wide Web offers a lot of data and useful information to companies wanting to sell seats of aircrafts, those companies invented algorithms to determine ticket prices. Etzioni, Knoblock, Tuchinda and Yates (2003) try to give an answer to the algorithmic setting of ticket prices by a pilot study in designing algorithm that predicts the price determination. With the knowledge of price determination, this algorithm should help you save a lot of money by the correct timing of purchasing your airfares.
By collecting data from a major travel web site, the scientists could find patterns in the changes of prices. For this research, the data of very similar trips were registered during 41 days. Etzioni et al. found that the prices of airfares can change up to seven times a day. These changes are separated into two dependent and independent changes. Dependent changes stems from the adjustment of prices by airlines to maximize their overall revenues. These changes happen in similar flights and occur once or twice a day. The independent changes probably result from change in seat availability.
The algorithm was a pilot research and therefore missed several key variables, such as the available seats in the airplane, to determine prices. However, Hamlet (the name of the algorithm) did surprisingly well in predicting prices. When following the advices of the pilot algorithm 61.8% of the possible savings was achieved. Having in mind these results, only from a pilot study, the potential of price mining algorithms to save consumers money is demonstrated.
Image source: Flickr / Kitty Terwolbeck