Prior Specification in Bayesian Model Averaging: An application to Economic Growth
Journal Title: Annals. Computer Science Series - Year 2018, Vol 16, Issue 1
Abstract
Some recent cross-country cross-sectional analyses have employed Bayesian Model Averaging to tackle the issue of model uncertainty. Bayesian model averaging has become an important tool in empirical settings with large numbers of potential regressors and relatively limited number of observations. We examine the effect of a variety of prior assumptions on the inference, posterior inclusion probabilities of regressors and on predictive performance. Bayesian model averaging (BMA) has become a widely accepted way of accounting for model uncertainty in regression models. However, to implement BMA, a prior is usually specified in two parts: prior for the regression parameters and prior over the model space. Hence, the choice of prior specification becomes paramount in Bayesian inference, unfortunately, in practice, most Bayesian analyses are performed with the so-called non-informative priors (i.e. priors constructed by some formal rule). The arbitrariness in the choice of prior or choosing inappropriate priors often lead to badly behaved posteriors. It is therefore imperative to study the effect of choice of priors in Bayesian model averaging. Six candidate parameter priors namely, Unit information prior (UIP), Risk inflation criterion (RIC), Bayesian Risk Inflation criterion (BRIC), Hannan-Quinn criterion (HQ), Empirical Bayes (EBL) and hyper-g and three model priors: uniform, beta-binomial and binomial were examined in this study. The performances of the resulting eighteen cases were judged using posterior inference, posterior inclusion probabilities of regressors and predictive performance. Analyses were carried out using datasets with 8-potential drivers of growth for 126 countries from 2010 to 2014. Finally, our analysis shows that the EBL parameter prior with random model prior robustly identifies far more growth determinants than other priors.
Authors and Affiliations
Tayo P. OGUNDUNMADE, Adedayo A. ADEPOJU
Keywords
Related Articles
- EP ID EP540182
- DOI -
- Views 82
- Downloads 0
An Exploratory Study of the Critical Factors Affecting the Acceptability of Automated Teller Machine (ATM) in Nigeria
This paper uses the Technology Acceptance Model (TAM) as a basis for studying critical factors that affects the acceptability of Automated Teller Machine (ATM) in Nigeria. Questionnaire approach was used with the respond...
Application Of Dimensionality Reduction On Classification Of Colon Cancer Using Ica And K-Nn Algorithm
Several sectors including engineering, health, academics and so on deals with very large number of information and few specimens. This highlight the need of a technique to improve data accuracy in order to enable profess...
Developing Predictive Models using Typical Machine Learning and Computational Techniques
This study investigates the accuracy of developing predictive models using machine learning techniques. The machine learning techniques considered in this study include artificial neural network (ANN) and Kalman filter a...
Computer applications in clinical psychology
The computer-assisted analysis is not currently a novelty, but a necessity in all areas of psychology. A number of studies that examine the limits of the computer assisted and analyzed interpretations, also its advantage...
Deep Learning Based Sentiment Analysis for Recommender System
Recommender Systems are used to predict and recommend products based on user preferences. Recommender System is a program that makes use of sentiment analysis technique. Sentiment analysis is one of the popular technique...