top of page

Bayesian Networks for information synthesis in complex systems

“We have plenty of data, now we need to know what to do with it.”

 

Many organisations around the world have large datasets distributed across many locations and formats. The sheer speed of data collection makes traditional data sharing and storage techniques extremely difficult, preventing information from being unlocked from data. The problem of increasingly complex and expensive methods of dealing with data has been addressed independently by different researchers for the last decade. However, as the various methods of dealing with large datasets have converged, researchers have formed a multi-disciplinary approach referred to as Big Data.

 

Nowhere is this problem more apparent than in airports, where multiple organisations work together to monitor and collect data on thousands of passengers a day. Information about passengers, their luggage and their movements is collected by these organisations using legacy systems built up over decades. The amorphous nature of data collection and storage across the airport makes traditional database-linking approaches extremely complicated.

 

My PhD thesis explored the use of Bayesian Networks to combine the information from all the organisations operating in the airport in a statistical model. Working with national level stakeholders including the Australian Customs and Border Protection Service, Department of Agriculture, Fisheries and Forestry, Brisbane International Airport and airlines, I began by mapping the data collection activities throughout the airport and determining which data were important to determining system performance. Based only on existing data, I created a hybrid Bayesian Network and queuing model that allows government regulators, airport managers and airlines to predict the number of passengers likely to be in the airport at any given time as well as the likely type of passengers (such as their language and reason for travel) and their processing requirements. This model has been adopted for use in Brisbane International Airport by Australian Customs, and has become the basis of a new Australian Research Council Linkage Project.

bottom of page