AER: Insights · 2022 · Arman Khachiyan、Anthony Thomas、Huye Zhou、Gordon Hanson、Alex Cloninger、Tajana Rosing
中文摘要
我们将深度学习应用于日间卫星影像，以在美国数据中预测高空间分辨率下的收入与人口变化。对于边长分别为 1.2 公里和 2.4 公里的网格单元（美国县域的平均边长为 51.9 公里），我们的模型预测在水平值上达到 0.85 至 0.91 的 R²，远超现有模型的精度；在十年变化量上达到 0.32 至 0.46 的 R²，这在文献中尚无对应结果，且比常用的夜间灯光数据大 3 至 4 倍。我们的网络在分析局部冲击方面具有广泛的应用价值。（JEL C45, R11, R23）
Abstract
We apply deep learning to daytime satellite imagery to predict changes in income and population at high spatial resolution in US data. For grid cells with lateral dimensions of 1.2 km and 2.4 km (where the average US county has dimension of 51.9 km), our model predictions achieve R 2 values of 0.85 to 0.91 in levels, which far exceed the accuracy of existing models, and 0.32 to 0.46 in decadal changes, which have no counterpart in the literature and are 3–4 times larger than for commonly used nighttime lights. Our network has wide application for analyzing localized shocks. (JEL C45, R11, R23)