A5010627600- Advanced Statistical Methods and Models
- Statistical Methods and Inference
- Control Systems and Identification
- Statistical Methods and Bayesian Inference
- Fault Detection and Control Systems
- Bayesian Methods and Mixture Models
- Structural Health Monitoring Techniques
- Forecasting Techniques and Applications
- Statistical Distribution Estimation and Applications
- Influenza Virus Research Studies
- Advanced Adaptive Filtering Techniques
- Neural Networks and Applications
- Complex Systems and Time Series Analysis
- Advanced Statistical Process Monitoring
- Financial Risk and Volatility Modeling
- Bayesian Modeling and Causal Inference
- Sensor Technology and Measurement Systems
- Image and Signal Denoising Methods
- Anomaly Detection Techniques and Applications
- Data Analysis with R
- Architecture and Computational Design
- Greenhouse Technology and Climate Control
- HIV/AIDS Impact and Responses
- Advanced Manufacturing and Logistics Optimization
- Simulation Techniques and Applications
Roma Tre University
2004-2022
Sapienza University of Rome
1998
Istituto Nazionale di Statistica
1992
Often the goal of model selection is to choose a for future prediction, and it natural measure accuracy prediction by squared error loss. Under Bayesian approach, commonly perceived that optimal predictive with highest posterior probability, but this not necessarily case. In paper we show that, among normal linear models, often median probability model, which defined as consisting those variables have overall greater than or equal 1/2 being in model. The differs from
The median probability model (MPM) (Barbieri and Berger, 2004) is defined as the consisting of those variables whose marginal posterior inclusion at least 0.5. MPM rule yields best single for prediction in orthogonal nested correlated designs. This result was originally conceived under a specific class priors, such point mass mixtures non-informative g-type priors. rule, however, has become so very popular that it now being deployed wider variety priors designs, where properties are not yet...
The problem of estimating the parameters a model for bidimensional data made up by linear combination damped two-dimensional sinusoids is considered. Frequencies, amplitudes, phases, and damping factors are estimated applying generalization monodimensional Prony's method to spatial data. This procedure finds desired quantities after an autoregressive fitting data, polynomial rooting, least-squares solution. models involved have particular nature that simplifies analysis. In fact, their...
This paper considers point null hypothesis testing when the sampling distribution belongs to a particular class, defined in Gleser & Hwang (1987). We discuss drawbacks of frequentist and likelihood solutions we show how proper Bayesian analysis encounters relatively similar difficulties. explore performance several noninformative approaches testing, namely asymptotic approximations Bayes factors default factors. argue that Fieller's problem choice 'correct' prior show, lesser extent,...
A Bayesian approach is adopted to the analysis of autoregressive time series subject outliers. Additive and innovational outliers are considered as particular cases a mixed generating model, which allows one handle situations in there may be an unknown number type. In paper exact form likelihood function used stationarity model enforced. The computational problems solved using version single component Metropolis-Hastings algorithm. method proposed obtain all posterior summaries interest...