Sign In

Communications of the ACM

ACM TechNews

New Research Accurately Predicts Australian Wheat Harvests Months in Advance


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
A field of wheat.

Research by the U.S. National Science Foundation found that machine learning techniques could anticipate the size of Australian wheat yields as long as two months before the end of the growing season.

Credit: Bluemoose

A U.S. National Science Foundation-funded project found that machine learning techniques could anticipate Australian wheat yields two months before the end of the growing season.

The University of Illinois' Kaiyu Guan said, "We tested various machine learning approaches and integrated large-scale climate and satellite data to come up with a reliable and accurate prediction of wheat production for the whole of Australia."

Guan noted the satellite data provided additional information which significantly refined wheat harvest prediction.

The researchers applied this information to forecast yield, with about 75% accuracy two months prior to crop maturation.

From National Science Foundation
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account