EURO 2024 Copenhagen
Abstract Submission

EURO-Online login

613. Improving the estimation of production functions through machine learning: a gradient boosting approach

Invited abstract in session MB-48: DEA and Machine Learning, stream Data Envelopment Analysis and its Application.

Monday, 10:30-12:00
Room: 60 (building: 324)

Authors (first author is the speaker)

1. María D. Guillén
Center of Operation Research, University Miguel Hernandez of Elche
2. Juan Aparicio
Center of Operations Research, Miguel Hernandez University of Elche

Abstract

Making accurate predictions of the true production frontier is critical for reliable efficiency analysis. However, traditional deterministic non-parametric methods like Free Disposal Hull (FDH) or Data Envelopment Analysis (DEA) provide approximations of the production frontier that suffer from overfitting, systematically underestimating firms' inefficiency and yielding inaccurate predictions of the output. In this work, we propose a new approach that, following the machine learning paradigm, provides a more accurate prediction of the underlying true production frontier by adapting the Gradient Tree Boosting algorithm to the production context. We prove that the new estimator satisfies certain required regulatory conditions such as envelopment of data and monotonicity. The performance of the new models is evaluated through a computation experience that shows the outperformance of the new approach in terms of mean squared error and bias in relation to the standard techniques. Moreover, we show how to calculate different efficiency measures using the estimator determined through the new algorithm. Nevertheless, from a computational point of view, the new approach presents thousands of decision variables, making it computationally complex to solve. To tackle this problem, we also propose and check a heuristic approximation for the exact measures.

Keywords

Status: accepted

Back to the list of papers