A1033
Title: Bayesian nonparametric machine learning approach for efficiency analysis
Authors: Zheng Wei - Texas A&M University (United States) [presenting]
Huiyan Sang - Texas A\&M University (United States)
Nene Coulibaly - Texas A and M University - Corpus Christi (United States)
Abstract: The stochastic frontier model is widely used in economics, finance, and management to estimate the production function and efficiency of a firm or industry. In the current literature on stochastic frontier analysis, parametric forms of the production function, such as Cobb-Douglas and translog, are often assumed a priori without validation, which may suffer from model misspecification and lead to biased efficiency estimates. To address this issue, a new stochastic frontier model built upon a monotone-constrained nonparametric production function is proposed via an extension of the monotone Bayesian Additive Regression Tree (MBART) framework, which allows for greater flexibility in modelling the production function with uncertainty measure while accounting for complex relationships between high dimensional data and variable selection. The performance of the proposed model was illustrated through simulation studies and a real data application.
