Search results for "Inference"
showing 10 items of 478 documents
Sampling properties of the Bayesian posterior mean with an application to WALS estimation
2022
Many statistical and econometric learning methods rely on Bayesian ideas, often applied or reinterpreted in a frequentist setting. Two leading examples are shrinkage estimators and model averaging estimators, such as weighted-average least squares (WALS). In many instances, the accuracy of these learning methods in repeated samples is assessed using the variance of the posterior distribution of the parameters of interest given the data. This may be permissible when the sample size is large because, under the conditions of the Bernstein--von Mises theorem, the posterior variance agrees asymptotically with the frequentist variance. In finite samples, however, things are less clear. In this pa…
Japan's FDI drivers in a time of financial uncertainty. New evidence based on Bayesian Model Averaging
2021
En este artículo analizamos los determinantes del stock de FDI saliente de Japón para el período 1996–2017. Este período es especialmente relevante ya que abarca un proceso de creciente globalización económica y dos crisis financieras. Para ello, consideramos un amplio conjunto de variables candidatas basadas en la teoría, así como en análisis empíricos previos. Nuestra muestra incluye un total de 27 países anfitriones. Seleccionamos las covariables utilizando una metodología basada en datos, el análisis Bayesian Model Averaging (BMA). Además, también analizamos si estos determinantes cambian según el grado de desarrollo (emergentes vs desarrollados) o las áreas geográficas (UE vs Asia Orie…
Unawareness and Partitional Information Structures
1999
Abstract We claim first that simple uncertainty is not an adequate model of a subject's ignorance, because a major component of it is the inability to give a complete description of the states of the world, and we provide a formal model of unawareness. In Modica and Rustichini (1994) we showed a difficulty in the project, namely that without weakening of the inference rules of the logic one would face the unpleasant alternative between full awareness and full unawareness. In this paper we study a logical system where non full awareness is possible, and prove that a satisfactory solution to the problem can be found by introducing limited reasoning ability of the subject. A determination theo…
The Simpson paradox of school grading in Italy
2009
Abstract Data from the 2003 OECD-PISA Survey for Italy reveal a striking difference in the relationship between students’ competence (as measured by PISA score in Mathematics) and school grades across regions: a competence level granting bare sufficiency in the North yields excellence grades in the South. This has spurred a lively debate on education policy in the country, based on the inference drawn from this evidence that grading practices are excessively different in the two areas. We show in this note that this inference overlooks a Simpson paradox hidden in the data. After a more careful analysis, the above inference is seen to be wrong. The crucial omitted variable is the school-leve…
An entropy-based machine learning algorithm for combining macroeconomic forecasts
2019
This paper applies a Machine Learning approach with the aim of providing a single aggregated prediction from a set of individual predictions. Departing from the well-known maximum-entropy inference methodology, a new factor capturing the distance between the true and the estimated aggregated predictions presents a new problem. Algorithms such as ridge, lasso or elastic net help in finding a new methodology to tackle this issue. We carry out a simulation study to evaluate the performance of such a procedure and apply it in order to forecast and measure predictive ability using a dataset of predictions on Spanish gross domestic product.
Broken rotor bars detection via Park's vector approach based on ANFIS
2014
Many attempts have been made on fault diagnosis of induction motors based on frequency and time domain analysis of stator current. In this paper, first the Park's vector transformation and frequency analysis for fault detection of induction motors are introduced. Then a smart approach using Adaptive Neuro Fuzzy Inference System (ANFIS) is proposed. This approach uses the time domain features derived from the Park's vector transformation of stator current. By the proposed method, a partial break including 5 mm crack on a bar, one broken bar and two broken bars using experimental data are investigated. It will be shown that features derived from Park's vector compared to features obtained fro…
Adaptive neural-fuzzy inference system based method to modeling of vehicle crash
2013
Various areas of research need to be considered in order to establish a mathematical model of a vehicle crash. To enhance the modeling process, a novel ANFIS-based approach to reconstruct behavior of impacting vehicles is presented in this paper. Kinematics of center of gravity (COG) a vehicle involved in an oblique barrier collision is reproduced by application of a five-layered ANFIS structure. Then, the same ANFIS system is used to simulate a different collision type than the one which was used in the training stage. The points of interests are selected to be the locations of accelerometers mounting. The accuracy of the proposed method is evaluated by the comparative analysis with the re…
Evaluation of statistical sampling for the assessment of residential consumption totals in water distribution networks
2014
The paper provides insights into stratified sampling, a standard statistical technique that may be employed to assess domestic water use in water distribution networks. The basic idea is to use only a few meters to provide inference on the total water consumption of a network or of a district metered area through the knowledge of some additional stratification variables, such as household typology, size and occupants number. Since any sampling procedure assumes that the variance of the variable at stake is known, either a suitable amount of past consumption data is necessary, or a specific preliminary survey must be carried out, in order to define the sampling plan. An application with real…
Comparison of different uncertainty techniques in urban stormwater quantity and quality modelling
2011
Abstract Urban drainage models are important tools used by both practitioners and scientists in the field of stormwater management. These models are often conceptual and usually require calibration using local datasets. The quantification of the uncertainty associated with the models is a must, although it is rarely practiced. The International Working Group on Data and Models, which works under the IWA/IAHR Joint Committee on Urban Drainage, has been working on the development of a framework for defining and assessing uncertainties in the field of urban drainage modelling. A part of that work is the assessment and comparison of different techniques generally used in the uncertainty assessm…
A fuzzy logic approach to modeling a vehicle crash test
2013
Published version of an article in the journal: Central European Journal of Engineering. Also available from the publisher at: http://dx.doi.org/10.2478/s13531-012-0032-2 This paper presents an application of fuzzy approach to vehicle crash modeling. A typical vehicle to pole collision is described and kinematics of a car involved in this type of crash event is thoroughly characterized. The basics of fuzzy set theory and modeling principles based on fuzzy logic approach are presented. In particular, exceptional attention is paid to explain the methodology of creation of a fuzzy model of a vehicle collision. Furthermore, the simulation results are presented and compared to the original vehic…