Efficiency measurement of research groups using Data Envelopment Analysis and Bayesian networks

Applications of non-parametric frontier production methods such as Data Envelopment Analysis (DEA) have gained popularity and recognition in scientometrics. DEA seems to be a useful method to assess the efficiency of research units in different fields and disciplines. However, DEA results give only a synthetic measurement that does not expose the multiple relationships between scientific production variables by discipline. Although some papers mention the need for studies by discipline, they do not show how to take those differences into account in the analysis. Some studies tend to homogenize the behaviour of different practice communities. In this paper we propose a framework to make inferences about DEA efficiencies, recognizing the underlying relationships between production variables and efficiency by discipline, using Bayesian Network (BN) analysis. Two different DEA extensions are applied to calculate the efficiency of research groups: one called CCRO and the other Cross Efficiency (CE). A BN model is proposed as a method to analyze the results obtained from DEA. BNs allow us to recognize peculiarities of each discipline in terms of scientific production and the efficiency frontier. Besides, BNs provide the possibility for a manager to propose what-if scenarios based on the relations found.

[1]  R MaríaIsabelRestrepo,et al.  Clasificación de grupos de investigación colombianos aplicando análisis envolvente de datos , 2007 .

[2]  Ronald Rousseau,et al.  Data envelopment analysis as a tool for constructing scientometric indicators , 1997, Scientometrics.

[3]  Sune Lehmann,et al.  A quantitative analysis of indicators of scientific performance , 2008, Scientometrics.

[4]  X. Polanco,et al.  Using a compound approach based on elaborated neural network for Webometrics: An example issued from the EICSTES project , 2004, Scientometrics.

[5]  W. Cooper,et al.  Data Envelopment Analysis: A Comprehensive Text with Models, Applications, References and DEA-Solver Software , 1999 .

[6]  A. U.S.,et al.  Measuring the efficiency of decision making units , 2003 .

[7]  Gregory F. Cooper,et al.  A Bayesian method for the induction of probabilistic networks from data , 1992, Machine Learning.

[8]  J. Pearl.  Statistics and causal inference: A review , 2003 .

[9]  P. Vanden Abeele,et al.  On research efficiency: A micro-analysis of Dutch university research in Economics and Business Management , 2005 .

[10]  J. Pearl.  Causality: Models, Reasoning and Inference , 2000 .

[11]  David Heckerman,et al.  A Tutorial on Learning with Bayesian Networks , 1999, Innovations in Bayesian Networks.

[12]  Nir Friedman,et al.  Being Bayesian About Network Structure. A Bayesian Approach to Structure Discovery in Bayesian Networks , 2004, Machine Learning.

[13]  Joe Zhu,et al.  Data Envelopment Analysis , 2007 .

[14]  Guan Jiancheng,et al.  Evaluation and interpretation of knowledge production efficiency , 2004 .

[15]  Yongtae Park,et al.  The impact of R&D collaboration on innovative performance in Korea: A Bayesian network approach , 2008, Scientometrics.

[16]  Ronald Rousseau,et al.  The scientific wealth of European nations: Taking effectiveness into account , 1998, Scientometrics.

[17]  Jiancheng Guan,et al.  Evaluation and interpretation of knowledge production efficiency , 2004, Scientometrics.

[18]  Henk F. Moed,et al.  Handbook of Quantitative Science and Technology Research , 2005 .

[19]  Rodney H. Green,et al.  Efficiency and Cross-efficiency in DEA: Derivations, Meanings and Uses , 1994 .

[20]  Jyrki Wallenius,et al.  Value efficiency analysis of academic research , 1998, Eur. J. Oper. Res..

[21]  Wei Meng,et al.  Efficiency evaluation of basic research in China , 2006, Scientometrics.

[22]  B. M. Gupta,et al.  Assessment of impact of AICTE funding on R&D and educational development , 2005, Scientometrics.

[23]  Jie Cheng,et al.  Learning Bayesian Networks from Data: An Efficient Approach Based on Information Theory , 1999 .

[24]  Léopold Simar,et al.  Advanced indicators of productivity of universitiesAn application of robust nonparametric methods to Italian data , 2006, Scientometrics.

[25]  Loet Leydesdorff,et al.  Knowledge representations, Bayesian inferences and empirical science studies , 1992 .

[26]  Geraldo da Silva e Souza,et al.  Technical efficiency of production in agricultural research , 1999, Scientometrics.

[27]  Richard H. Silkman,et al.  Measuring efficiency : an assessment of data envelopment analysis , 1986 .

[28]  Richard Scheines,et al.  Discovery Algorithms for Causally Sufficient Structures , 1993 .

[29]  T. Sexton,et al.  Data Envelopment Analysis: Critique and Extensions , 1986 .

[30]  Cinzia Daraio,et al.  Econometric Approaches to the Analysis of Productivity of R&D Systems , 2004 .

[31]  P. Spirtes,et al.  Causation, prediction, and search , 1993 .