Graph Sparsification for GCN Towards Optimal Crop Yield Predictions

TL;DR

This paper introduces a graph sparsification technique based on the Fiedler number to improve the efficiency of GCNs in crop yield prediction, maintaining accuracy while reducing computational complexity.

Contribution

It presents a novel Fiedler-based graph sparsification method that effectively reduces graph density for GCNs in agricultural yield prediction tasks.

Findings

Fiedler-based sparsification achieves comparable accuracy to dense graphs.

The method reduces training and inference time significantly.

Compared to other schemes, it maintains better GCN performance.

Abstract

In agronomics, predicting crop yield at a per field/county granularity is important for farmers to minimize uncertainty and plan seeding for the next crop cycle. While state-of-the-art prediction techniques employ graph convolutional nets (GCN) to predict future crop yields given relevant features and crop yields of previous years, a dense underlying graph kernel requires long training and execution time. In this paper, we propose a graph sparsification method based on the Fiedler number to remove edges from a complete graph kernel, in order to lower the complexity of GCN training/execution. Specifically, we first show that greedily removing an edge at a time that induces the minimal change in the second eigenvalue leads to a sparse graph with good GCN performance. We then propose a fast method to choose an edge for removal per iteration based on an eigenvalue perturbation theorem. Experiments show that our Fiedler-based method produces a sparse graph with good GCN performance compared to other graph sparsification schemes in crop yield prediction.