Online Program

Return to main conference page
Friday, February 15
Fri, Feb 15, 11:00 AM - 12:30 PM
Magazine
Special Applications of Statistical Software

Computational Advances in the Production of Compact Letter Displays (303753)

View Presentation View Presentation

*John M. Ennis, The Institute for Perception 
William J. Russ, The Institute for Perception 

Keywords: Multiple Statistical Comparisons, Letter Displays, Linear Programming, Computational Efficiency

Letter displays are valuable for representing the results of a large number of multiple statistical comparisons in a single table. This value notwithstanding, letter displays often include numerous unnecessary letter assignments. In this presentation we detail recent advances for determining which combinations of letter assignments are unnecessary. In particular, we present new algorithms which employ linear programming techniques from operations research to find compact letter displays that are minimal with respect to either: a) the number of distinct letters or b) the number of total letters used by the display. We then compare these algorithms with previously state-of-the-art algorithms for compact letter display production to show that these linear programming based minimizations offer a dramatic improvement in computational efficiency.