|
Abstract:
|
Conferences are a vital part of any profession. They provide an opportunity for people of similar interests to come together and share ideas and best practice. They are important, but they are also complicated to plan. Consider for a moment the Joint Statistical Meetings - the largest annual gathering of statisticians. It runs for six days, welcomes more than 6500 delegates and involves more than 600 different sessions. Arranging all the sessions with their associated documents is a massive undertaking for the organizing committee, taking months to develop a comprehensive and cohesive program. What if the committee's work could be streamlined by algorithms in a matter of hours? We describe an experimental approach to the problem of automated conference programming, combining text-mining tools with aspects of computer science known as discrete optimization. An example using text documents to optimize conference planning is illustrated.
|
