Research: Now Blooming: Digital Models

Now Blooming: Digital Models
Washington Post (03/29/08) P. B1; Ruane, Michael E.

Virginia Tech master's students Vidhya Dass and Elizabeth Brennan are using artificial neural networks, evolutionary computations, the Arrhenius equation, linear regression, and fuzzy logic to predict when Washington, D.C.'s cherry trees will bloom. Dass and Brennan wanted to see if a computer model could do as well or better than the National Park Service's seasoned horticulturalist, who analyzes such factors as early flowering elms, maples, and cornelian cherry dogwoods, as well as the weather and other recurring clues. An accurate computer model could make it easier for officials to plan the National Cherry Blossom Festival and for tourists to plan visits. "We hoped to create a model that would allow the best prediction with the minimum amount of input," Brennan says. Dass and Brennan say they focused most of their efforts on computational intelligence and essentially tried to mimic a human brain. The students point out that computer modeling is widely used in to predict soybean flowering, corn yields, and aspects of tomato and lettuce farming. They used past peak dates and previously recorded data to see which computer models were the most accurate. The most accurate models matched past peak dates to within a few days, and some models were as much as three days closer to the peak bloom date than the park service's prediction for that year.
Click Here to View Full Article