CMStatistics 2017: Start Registration
View Submission - CMStatistics
B1044
Title: Inference for nonparametric productivity networks: A pseudo-likelihood approach Authors:  Cinzia Daraio - University of Rome La Sapienza (Italy) [presenting]
Rolf Fare - Oregon State University (United States)
Shawna Grosskopf - Oregon State University (United States)
Maria Grazia Izzo - University of Rome La Sapienza (Italy)
Luca Leuzzi - CNR National Research Council (Italy)
Giancarlo Ruocco - University of Rome La Sapienza (Italy)
Moriah Bostian - Lewis and Clark College (United States)
Abstract: There is a rich literature on the nonparametric estimation of efficiency (Data Envelopment Analysis) based on networks (Network DEA) which typically analyzes the networks in a descriptive rather than statistical framework. The goal is fill this gap by developing a new more general framework for modeling the production process to include estimation of production functions, information theoretic approaches to econometrics, machine learning and statistical inference from the physics of complex systems. We combine this general model with Georgescu Roegen's model of flows and funds according to a previous production function. The proposed statistical approach is to reconstruct the network's topology for nonparametric frontier models, based on recent pseudo-likelihood techniques, allowing us to infer the network topology in a Bayesian framework. We illustrate this approach with an application to assess the scientific productivity of world countries at a disciplinary level.