South Dakota Mines students are building computer models to predict cattle markets. The software even accounts for anomalies like droughts.
The students are relying on work originally done decades ago. In the early 1990s, two Mines students, Ron Ragsdale and Todd Gagne, developed the software.
Gagne says it worked for Ragsdale, a mathematician and rancher.
"We made him some tools that said when the actual price of the animal gets out of whack. Then, he would buy these future contracts and ride them to fruition. And because they would have to normalize, he made money," Gagne says.
Ragsdale and Gagne continued to use and tweak the software — until they saw an opportunity with machine learning.
In 2021, Gagne shared the software with Mines students. He also brought historical data for both cattle and corn.
"We started to use machine learning and saying, 'Can you predict — based on all this historical data — and start to do some predictive analytics on where these markets are going to go?'" Gagne says.
The team's next focus is determining which of the 187 variables accounted for by the software carries the most weight.
Once the team feels the software is robust enough, they plan to commercialize it.
Learn about this research by clicking here.
