Estimating Parameters of Structural Models Using Neural Networks

46 Pages Posted: 9 Dec 2019 Last revised: 23 Apr 2021

See all articles by Yanhao 'Max' Wei

Yanhao 'Max' Wei

University of Southern California - Marshall School of Business

Zhenling Jiang

University of Pennsylvania - The Wharton School

Date Written: April 23, 2020

Abstract

Machine learning tools such as neural networks are increasingly applied in marketing and economics to learn complex relations in data. The learned relations allow machines to perform various tasks, such as recognizing objects from images or recognizing emotions from speech. This paper explores using a neural net to learn the relation between data (moments) and the parameter values of a structural economic model, so that it can “recognize,” or estimate, these parameter values from the data (moments). We train the neural net with the datasets generated by the structural model under different parameter values. The neural net can be trained to give not only the point estimates of parameters but also their statistical accuracy. We show this Neural Net Estimator (NNE) converges to meaningful and well-known limits when the number of training datasets is sufficiently large. NNE does not require computing integrals over the unobservables in the structural model. Thus, it is suitable for models where such integrals are costly in MLE/GMM. We benchmark NNE in two Monte Carlo studies. NNE is able to achieve high estimation accuracies under very light estimation costs.

Keywords: structural estimation, computational costs, neural networks, machine learning, market entry, choice models.

Suggested Citation

Wei, Yanhao and Jiang, Zhenling, Estimating Parameters of Structural Models Using Neural Networks (April 23, 2020). USC Marshall School of Business Research Paper, Available at SSRN: https://ssrn.com/abstract=3496098 or http://dx.doi.org/10.2139/ssrn.3496098

Yanhao Wei (Contact Author)

University of Southern California - Marshall School of Business ( email )

701 Exposition Blvd
Los Angeles, CA California 90089
United States

Zhenling Jiang

University of Pennsylvania - The Wharton School ( email )

3641 Locust Walk
Philadelphia, PA 19104-6365
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
347
Abstract Views
1,484
rank
108,314
PlumX Metrics