Conference Program

Predicting Crop-Specific Land Cover Using Transition Probabilities, Deep and Quantum-Inspired Neural Network Models (310067)

Claire Boryan, USDA National Agricultural Statistics Service 
Andrew Dau, USDA National Agricultural Statistics Service 
*Luca Sartore, National Institute of Statistical Sciences 
Patrick Willis, USDA National Agricultural Statistics Service 

Keywords: Categorical Prediction, Transition Probability, Deep Neural Networks, Crop Rotation Patterns

Traditional crop-rotation models suffer from sparsity issues due to the representation of categorical data as indicator variables or parametric spaces that exponentially increase with longer rotation patterns. In this paper, a quantum-inspired neural network model that has the potential of processing longer crop-rotation patterns over larger areas is proposed and compared to two established models.