Farms.com Home   News

Evaluating an A.I. Tool on Wisconsin Soybean and Corn Plots

Evaluating an A.I. Tool on Wisconsin Soybean and Corn Plots

By Shawn P. Conley

IN A BEAN POD:

  • Soybean A.I.-based cropping systems were, in general, successful in increasing yield and profit compared to typical systems
  • Across all locations, following soybean A.I. recommended systems would have increased mean yield by ~7 bu/ac and mean profit by ~$40/ac compared to typically used cropping systems
  • The potential of corn A.I.-based cropping systems to increase yield and profit was not clear
  • The corn A.I. tool recommended systems resulted in either increased or similar profit with typical systems by applying 19-223% lower nitrogen fertilizer rate

Algorithm-based decision making will likely play an important role in the coming years. Algorithms can capture and quantify complex relationships that can result in more informative decisions with greater probability of success (effectively increase profit) compared to current approaches. Evaluation of such tools in field conditions which involve unexpected and unmanageable yield adversities is important.

In this work, soybean A.I.-based cropping systems were in general successful to increase yield and profit compared to typical systems. The potential of corn A.I.-based cropping systems to increase yield and profit though was not clear. Additionally, Tar Spot was found and not treated at all three locations. This may have impacted the overall results of the experiment and suggest that the A.I. tool alone cannot account for in-season IPM decisions and should be paired with scouting or other management tools such as TarSpotter.

Source : wisc.edu

Trending Video

CFTC Chair Defends Enforcement

Video: CFTC Chair Defends Enforcement

This week, the House Agriculture Committee heard testimony from Michael Selig, the Chairman of the Commodity Futures Trading Commission. Many of the questions focused on farmer’s access to trading and pricing data, and enforcement of insider trading on prediction markets.