A0487
Title: Calibration of spatio-temporal forecasts from urban air pollution data with sparse recurrent neural networks
Authors: Matthew Bonas - University of Notre Dame (United States) [presenting]
Stefano Castruccio - University of Notre Dame (United States)
Abstract: Data collected from personal air quality monitors have become an increasingly valuable tool to complement existing public health monitoring systems in urban areas. The potential of using such `citizen science data' for automatic early warning systems is hampered by the lack of models able to capture the high resolution, nonlinear spatio-temporal features stemming from local emission sources such as traffic, residential heating and commercial activities. A machine learning approach si proposed to forecast high-frequency spatial fields which has two advantages from standard methods in time: 1) sparsity of the neural network via a spike-and-slab prior, and 2) a small parametric space. The introduction of stochastic neural networks generates additional uncertainty, and we propose a fast approach to ensure that the forecast is correctly assessed (calibration), both marginally and spatially. We focus on assessing exposure to urban air pollution in San Francisco, and our results suggest an improvement of over 30\% in the mean squared error over the standard time series approach with a calibrated forecast for up to 5 days.
