Some Cracks in the Empire of Chance (Flaws in the Foundations of Reliability)

Making quantified statements about the uncertainty associated with the lifelength of an item is one of the most fundamental tasks of reliability assessment. Most practitioners routinely do this using one of the several available statistical techniques. The purpose of this paper is two-fold. The first is to give the user an overview of the key tenets of two of the most commonly used parametric approaches. The second is to point out that these commonly used approaches involve strategies that are either ad hoc, or are in violation of some of the underlying tenets. A method that is devoid of logical flaws can be proposed, but this method is difficult to implement. The user must therefore resign to using that technique against which the fewest objections can be hurled.     

Conditioning as disintegration

Conditional probability distributions seem to have a bad reputation when it comes to rigorous treatment of conditioning. Technical arguments are published as manipulations of Radon–Nikodym derivatives, although we all secretly perform heuristic calculations using elementary definitions of conditional probabilities. In print, measurability and averaging properties substitute for intuitive ideas about random variables behaving like constants given particular conditioning information.
One way to engage in rigorous, guilt-free manipulation of conditional distributions is to treat them as disintegrating measures—families of probability measures concentrating on the level sets of a conditioning statistic. In this paper we present a little theory and a range of examples—from EM algorithms and the Neyman factorization, through Bayes theory and marginalization paradoxes—to suggest that disintegrations have both intuitive appeal and the rigor needed for many problems in mathematical statistics.     

Bayes方法在教学效度分析上应用与实现

本文提出了基于Bayes理论方法的多元数据融合方法,给出了该方法在教学改革中教学质量综合评估中的应用,并结合体育课内容制定的实例证明了该方法的有效性。     

Modelling of repairable systems with various degrees of repair

The availability of a stochastic repairable system depends on the failure behaviour and on repair strategies. In this paper, we deal with a general repair model for a system using auxiliary counting processes and corresponding intensities which include various degrees of repair (between minimal repair and perfect repair). For determining the model parameters we need estimators depending on failure times and repair times: maximum likelihood (ML) estimator and Bayes estimators are considered. Special results are obtained by the use of Weibull-type intensities and random observation times.     

Bayesian assessment of dimensionality in reduced rank regression

We consider Bayesian inference about the dimensionality in the multivariate reduced rank regression framework, which encompasses several models such as MANOVA, factor analysis and cointegration models for multiple time series. The fractional Bayes approach is used to derive a closed form approximation to the posterior distribution of the dimensionality and some asymptotic properties of the approximation are proved. Finite sample properties are studied by simulation and the method is applied to growth curve data and cointegrated multivariate time series.     

Some Variants of Constrained Estimation in Finite Population Sampling

This is a review article that unifies several important examples using constrained optimisation techniques. The basic tools are three simple mathematical optimisation results subject to certain constraints. Applications include calibration, benchmarking in small area estimation and imputation. A final illustration is constrained optimisation under a general divergence loss.     

The role of text-extracted investor sentiment in Chinese stock price prediction with the enhancement of deep learning

Whether investor sentiment affects stock prices is an issue of long-standing interest for economists. We conduct a comprehensive study of the predictability of investor sentiment, which is measured directly by extracting expectations from online user-generated content (UGC) on the stock message board of Eastmoney.com in the Chinese stock market. We consider the influential factors in prediction, including the selections of different text classification algorithms, price forecasting models, time horizons, and information update schemes. Using comparisons of the long short-term memory (LSTM) model, logistic regression, support vector machine, and Naïve Bayes model, the results show that daily investor sentiment contains predictive information only for open prices, while the hourly sentiment has two hours of leading predictability for closing prices. Investors do update their expectations during trading hours. Moreover, our results reveal that advanced models, such as LSTM, can provide more predictive power with investor sentiment only if the inputs of a model contain predictive information.     

Objective methods for graphical structural learning

Graphical models are used for expressing conditional independence relationships among variables by the means of graphs, whose structure is typically unknown and must be inferred by the data at hand. We propose a theoretically sound Objective Bayes procedure for graphical model selection. Our method is based on the Expected-Posterior Prior and on the Power-Expected-Posterior Prior. We use as input of the proposed methodology a default improper prior and suggest computationally efficient approximations of Bayes factors and posterior odds. In a variety of simulated scenarios with varying number of nodes and sample sizes, we show that our method is highly competitive with, or better than, current benchmarks. We also discuss an application to protein-signaling data, which wieldy confirms existing results in the scientific literature.     

Pareto-undominated and socially-maximal equilibria in non-atomic games

This paper makes the observation that a finite Bayesian game with diffused and disparate private information can be conceived of as a large game with a non-atomic continuum of players. By using this observation as its methodological point of departure, it shows that (i) a Bayes–Nash equilibrium (BNE) exists in a finite Bayesian game with private information if and only if a Nash equilibrium exists in the induced large game, and (ii) both Pareto-undominated and socially-maximal BNE exist in finite Bayesian games with private information. In particular, it shows these results to be a direct consequence of results for a version of a large game re-modeled for situations where different players may have different action sets.