Grain Quality Predictive Modeling

Student

Abby Slater

Advisor

Daniel Boley

Abstract

In today’s modern marketplace, companies are locked in an arms race against each other, constantly vying for any type of competitive advantage they can muster. The agricultural industry is no different. When it comes to grain, intrinsic factors like odor and color, combine with extrinsic factors like moisture and foreign matter, to determine the overall quality of the grain that a company produces. As grain is transported--often down the Mississippi River--the quality of the grain deteriorates in transit due to incorrect moisture levels, the introduction of foreign objects, or overall damage to the grain. This project focuses on creating and implementing predictive models which utilize factors such as the temperature and humidity of the grain, as well as its ambient temperature, to predict grain quality upon reaching its destination. Linear regression techniques are used in modeling, and the results are evaluated using metrics such as the mean squared error (MSE), the root mean squared error (RMSE), and the mean absolute error (MAE). These models provide a crucial first step in creating a tool that would allow agricultural companies to input the origin data of the grain, and receive predicted destination grades for moisture, foregin material content, and grain damage.