2299. Measuring the Malmquist Productivity Index Incorporating Probabilistic Variations in Data
Invited abstract in session TB-60: DEA under uncertainty, stream Data Envelopment Analysis and its applications.
Tuesday, 10:30-12:00Room: Western LT
Authors (first author is the speaker)
| 1. | Yu Zhao
|
| School of Management, Department of Management, Tokyo University of Science |
Abstract
The Malmquist Productivity Index is a widely used tool for evaluating productivity growth. In this study, we address the probabilistic variations present in input-output data and propose a forest-based sampling approach to account for the statistical properties of efficiency scores when constructing the Malmquist Productivity Index. To capture the underlying probability distributions, we classify observed decision-making units (DMUs) into clusters and optimize information gain within each cluster using a Gaussian-based entropy function. Additionally, we discuss the decomposition of the Malmquist Productivity Index into catch-up and frontier-shift components, along with the construction of confidence intervals for these components. The effectiveness of the proposed method is demonstrated through both simulated and empirical datasets.
Keywords
- Data Envelopment Analysis
- Algorithms
Status: accepted
Back to the list of papers