Feb 22 2017 stat.AP
This paper presents an estimator for semiparametric models that uses a feed-forward neural network to fit the nonparametric component. Unlike many methodologies from the machine learning literature, this approach is suitable for longitudinal/panel data. It provides unbiased estimation of the parametric component of the model, with associated confidence intervals that have near-nominal coverage rates. It is further shown that this model and estimator nests a nonparametric heterogeneous treatment effects model and estimator, which can consistently estimate individualized treatment effects conditional on covariates. Simulations demonstrate (1) efficiency, (2) that parametric estimates are unbiased, and (3) coverage properties of estimated intervals. An application section demonstrates the method by predicting county-level corn yield using daily weather data from the period 1981-2015, along with parametric time trends representing technological change. The method is shown to out-perform linear methods such as OLS and ridge/lasso, as well as random forest. The procedures described in this paper are implemented in the R package panelNNET.