A Study on Graph-Structured Recurrent Neural Networks and Sparsification with Application to Epidemic Forecasting
This work addresses epidemic forecasting for public health applications, but it is incremental as it builds on existing methods with efficiency improvements.
The authors tackled epidemic forecasting on real-world health data using a graph-structured recurrent neural network, achieving state-of-the-art accuracy on the CDC benchmark and maintaining prediction accuracy while sparsifying the network to have 70% zero weights.
We study epidemic forecasting on real-world health data by a graph-structured recurrent neural network (GSRNN). We achieve state-of-the-art forecasting accuracy on the benchmark CDC dataset. To improve model efficiency, we sparsify the network weights via transformed-$\ell_1$ penalty and maintain prediction accuracy at the same level with 70% of the network weights being zero.