A Bayesian approach for predicting match outcomes: The 2006 (Association) Football World Cup

In this paper we propose a Bayesian methodology for predicting match outcomes. The methodology is illustrated on the 2006 Soccer World Cup. As prior information, we make use of the specialists’ opinions and the FIFA ratings. The method is applied to calculate the win, draw and loss probabilities at each match and also to simulate the whole competition in order to estimate classification probabilities in group stage and winning tournament chances for each team. The prediction capability of the proposed methodology is determined by the DeFinetti measure and by the percentage of correct forecasts.     

On Markov chain Monte Carlo Algorithms for Computing Conditional Expectations Based on Sufficient Statistics

Much work has focused on developing exact tests for the analysis of discrete data using log linear or logistic regression models. A parametric model is tested for a dataset by conditioning on the value of a sufficient statistic and determining the probability of obtaining another dataset as extreme or more extreme relative to the general model, where extremeness is determined by the value of a test statistic such as the chi-square or the log-likelihood ratio. Exact determination of these probabilities can be infeasible for high dimensional problems, and asymptotic approximations to them are often inaccurate when there are small data entries and/or there are many nuisance parameters. In these cases Monte Carlo methods can be used to estimate exact probabilities by randomly generating datasets (tables) that match the sufficient statistic of the original table. However, naive Monte Carlo methods produce tables that are usually far from matching the sufficient statistic. The Markov chain Monte Carlo method used in this work (the regression/attraction approach) uses attraction to concentrate the distribution around the set of tables that match the sufficient statistic, and uses regression to take advantage of information in tables that “almost” match. It is also more general than others in that it does not require the sufficient statistic to be linear, and it can be adapted to problems involving continuous variables. The method is applied to several high dimensional settings including four-way tables with a model of no four-way interaction, and a table of continuous data based on beta distributions. It is powerful enough to deal with the difficult problem of four-way tables and flexible enough to handle continuous data with a nonlinear sufficient statistic.     

LP-metric sensitivity analysis for single and multi-attribute decision analysis

Analyzing the sensitivity of decisions to probability estimation error in single and multi-attribute problems and to errors in estimating additive multi-attribute value models in multi-attribute problems is an integral part of decision analysis. This paper presents an intuitive and tractable approach to this sensitivity analysis. Here a decision is considered insensitive if: 1) the probabilities or multi-attribute weights required for any other alternative to become preferred are not close to the original estimated probabilities and weights, and 2) the rank order of states implied by the probabilities or the rank order of attributes implied by the additive multi-attribute weights must change for any other alternative to become preferred. The sensitivity analysis is conducted using straight forward linear programming models. An example is used to demonstrate their application.     

Non-Parametric Change-Point Estimation using String Matching Algorithms

Given the output of a data source taking values in a finite alphabet, we wish to estimate change-points, that is times when the statistical properties of the source change. Motivated by ideas of match lengths in information theory, we introduce a novel non-parametric estimator which we call CRECHE (CRossings Enumeration CHange Estimator). We present simulation evidence that this estimator performs well, both for simulated sources and for real data formed by concatenating text sources. For example, we show that we can accurately estimate the point at which a source changes from a Markov chain to an IID source with the same stationary distribution. Our estimator requires no assumptions about the form of the source distribution, and avoids the need to estimate its probabilities. Further, establishing a fluid limit and using martingale arguments.     

Random effects model for credit rating transitions

This paper proposes a random effects multinomial regression model to estimate transition probabilities of credit ratings. Unlike the previous studies on the rating transition, we applied a random effects model, which accommodates not only the environmental characteristics of the exposures of a rating but also the uncertainty not explained by such factors. The rating category specific factors such as retained earning and market equity are included in our proposed model. The random effects model provides less diagonally dominant matrix, where the transition probabilities are over-dispersed from the diagonal elements. Our study is expected to incorporate potential chances of rating transitions due to extra random variations.     

Some Consequences of a Three-Party Split in a UK General Election

The British electoral system is unique in Europe in being of the first-past-the-post variety. The apparent emergence of a strong third party renders any prediction exercise a good deal more difficult, although some political commentators appear oblivious to that fact. It would appear that a transition matrix approach is capable of providing the deeper insights needed to explore the consequences of alternative voting-behaviour patterns. Unfortunately data of this form are not currently collected, but it is possible to postulate that transition matrices of a particular form could be of interest to the three parties. By associating such transition probabilities with the 1983 results for each of the 633 mainland constituencies, one can derive interesting relationships between the number of seats secured by each party.A range of computer analyses was performed, and this article sets out some of the more interesting results, some of which came as something of a surprise.     

Numerical Approach for Assessing System Dynamic Availability Via Continuous Time Homogeneous Semi-Markov Processes

Continuous-time homogeneous semi-Markov processes (CTHSMP) are important stochastic tools to model reliability measures for systems whose future behavior is dependent on the current and next states occupied by the process as well as on sojourn times in these states. A method to solve the interval transition probabilities of CTHSMP consists of directly applying any general quadrature method to the N ² coupled integral equations which describe the future behavior of a CTHSMP, where N is the number of states. However, the major drawback of this approach is its considerable computational effort. In this work, it is proposed a new more efficient numerical approach for CTHSMPs described through either transition probabilities or transition rates. Rather than N ² coupled integral equations, the approach consists of solving only N coupled integral equations and N straightforward integrations. Two examples in the context of availability assessment are presented in order to validate the effectiveness of this method against the comparison with the results provided by the classical and Monte Carlo approaches. From these examples, it is shown that the proposed approach is significantly less time-consuming and has accuracy comparable to the method of N ² computational effort.     

Hidden heterogeneity in manpower systems: A Markov-switching model approach

In modeling manpower systems, it is of crucial importance to deal with heterogeneity. Until recently, manpower models are dealing with heterogeneity due to observable sources, neglecting heterogeneity due to latent sources. In this paper a two-step procedure is introduced. In the first step personnel groups homogeneous with respect to the transition probabilities are determined in a classical way by taking into account the observable sources of heterogeneity. In the second step heterogeneity caused by latent sources is handled. A multinomial Markov-switching manpower model is introduced that deals with heterogeneity due to latent sources for the internal flows as well as for the wastage flows. The model incorporates the mover-stayer principle. A re-estimation algorithm is presented to estimate the parameters of the Markov-switching manpower model. The switching approach offers a methodology to build a Markov model with personnel groups as states that are more homogeneous, and therefore can contribute to a better validity of the manpower model.     

Kolmogorov equations in fractional derivatives for the transition probabilities of continuous-time Markov processes

A family of one-dimensional continuous-time Markov processes is considered, for which the author has earlier determined the transition probabilities by directly solving the Kolmogorov–Chapman equation; these probabilities have the form of single integrals. Analogues of the first and second Kolmogorov equations for the family of processes under consideration are obtained by using a procedure for obtaining integro-differential equations describing Markov processes with discontinuous trajectories. These equations turn out to be equations in fractional derivatives. The results are based on an asymptotic analysis of the transition probability as the start and end times of the transition approach each other. This analysis implies that the trajectories of a given Markov process are divided into two classes, depending on the interval in which they start. Some of the trajectories decay during a short time interval with a certain probability, and others are generated with a certain probability.     

Rare-Event Simulation of Non-Markovian Queueing Networks Using a State-Dependent Change of Measure Determined Using Cross-Entropy

A method is described for the efficient estimation of small overflow probabilities in nonMarkovian queueing network models. The method uses importance sampling with a state-dependent change of measure, which is determined adaptively using the cross-entropy method, thus avoiding the need for a detailed mathematical analysis. Experiments show that the use of rescheduling is needed in order to get a significant simulation speedup, and that the method can be used to estimate overflow probabilities in a two-node tandem queue network model for which simulation using a state-independent change of measure does not work well.     

The Most Critical Path in a PERT Network

The classical PERT approach uses the path with the largest expected duration as the critical path to estimate the probability of completing a project by a given deadline. However, in general, such a path is not the ‘most critical’ path and does not provide the smallest estimate for the probability of completion time. This paper studies the ‘most critical path’ problem and formulates it as an optimal path problem in a deterministic network with a two-attribute fractional objective function. An exact solution approach is presented for the optimal path problem which also gives the solution to the most critical path problem. The illustrative examples as well as our computational results demonstrate that the proposed algorithm provides estimates for the probabilities of completion time that are much more accurate than those of the classical approach.     

The Conditional-Potts Clustering Model

This article presents a Bayesian kernel-based clustering method. The associated model arises as an embedding of the Potts density for class membership probabilities into an extended Bayesian model for joint data and class membership probabilities. The method may be seen as a principled extension of the super-paramagnetic clustering. The model depends on two parameters: the temperature and the kernel bandwidth. The clustering is obtained from the posterior marginal adjacency membership probabilities and does not depend on any particular value of the parameters. We elicit an informative prior based on random graph theory and kernel density estimation. A stochastic population Monte Carlo algorithm, based on parallel runs of the Wang–Landau algorithm, is developed to estimate the posterior adjacency membership probabilities and the parameter posterior. The convergence of the algorithm is also established. The method is applied to the whole human proteome to uncover human genes that share common evolutionary history. Our experiments and application show that good clustering results are obtained at many different values of the temperature and bandwidth parameters. Hence, instead of focusing on finding adequate values of the parameters, we advocate making clustering inference based on the study of the distribution of the posterior adjacency membership probabilities. This article has online supplementary material.     

Modelling match outcomes and decision support for setting a final innings target in test cricket

^** E-mail: p.a.scarf{at}salford.ac.uk This paper considers the problem of timing the declaration ofthe third innings in test cricket. Data on the outcomes of recenttest matches are analysed in order to develop simple decisionsupport tools. The first of these tools presents match outcomeprobabilities given the position of the match at a potentialdeclaration point. These probabilities are determined usinga multinomial logistic regression model that is fitted to thetest match data. This idea is then extended to consider progresstowards a declaration—match outcome probabilities areconsidered as a function of target aimed for and run-rate. Thedecision tools have been implemented on a spreadsheet and examplecalculations are presented. The modelling described has thepotential for practical use in test matches.     

Simultaneous Posterior Probability Statements From Monte Carlo Output

This article considers the problem of making simultaneous probability statements in multivariate inferential problems based on samples from a posterior distribution. The calculation of simultaneous credible bands is reviewed and—as an alternative—contour probabilities are proposed. These are defined as 1 minus the content of the highest posterior density region which just covers a certain point of interest. We discuss a Monte Carlo method to estimate contour probabilities and distinguish whether or not the functional form of the posterior density is available. In the latter case, an approach based on Rao-Blackwellization is proposed. We highlight that this new estimate has an important invariance property. We illustrate the performance of the different methods in three applications.     

Spanning Forests of Digraphs and Limiting Probabilities of Markov Chains

A maximum out forest of a digraph is its spanning subgraph that consists of disjoint diverging trees and has the maximum possible number of arcs. For an arbitrary weighted digraph, we consider a matrix of specific weights of maximum out forests and demonstrate how this matrix can be used to get a graph-theoretic interpretation for the limiting probabilities of Markov chains. For a special (nonclassical) correspondence between Markov chains and weighted digraphs, the matrix of Cesáro limiting transition probabilities of any finite homogeneous Markov chain coincides with the normalized matrix of maximum out forests of the corresponding digraphs. This provides a finite (combinatorial) method to calculate the limiting probabilities of Markov chains and thus their stationary distributions. On the other hand, the Markov chain technique provides the proofs to some statements about digraphs.     

Forecasting the partisan composition of the house as a function of seniority

This note examines the influence of seniority of the partisan composition of the House of Representatives. A model is developed to compute the re-election probabilities as a function of the levels of seniority. These estimates, together with the composition of the immediately preceding congress, are then used to estimate the composition of any congress and the results compared to the actual electoral outcomes. The end result is a partisan composition or the House independent of the short-term fluctuations created by issues relevant to the individual elections. In addition, measures of accuracy of the estimates are computed in order to assertain how much of the variation in the data is left to be explained by issue-oriented factors.     

Bayesian corporate bond pricing and credit default swap premium models for deriving default probabilities and recovery rates

This paper develops a Bayesian method by jointly formulating a corporate bond (CB) pricing model and credit default swap (CDS) premium pricing models to estimate the term structure of default probabilities and the recovery rate. These parameters are formulated by incorporating firm characteristics such as industry, credit rating and Balance Sheet/Profit and Loss information. A cross-sectional model valuing all given CB prices and CDS premiums is considered. The quantities derived are regarded as what market participants infer in forming CB prices and CDS premiums. We also develop a statistical significance test procedure without any distributional assumptions for the specified model. An empirical analysis is conducted using Japanese CB and CDS market data.     

Exact probabilities of correct classifications for uncorrelated repeated measurements from elliptically contoured distributions

Euclidean distance-based classification rules are derived within a certain nonclassical linear model approach and applied to elliptically contoured samples having a density generating function g. Then a geometric measure theoretical method to evaluate exact probabilities of correct classification for multivariate uncorrelated feature vectors is developed. When doing this one has to measure suitably defined sets with certain standardized measures. The geometric key point is that the intersection percentage functions of the areas under investigation coincide with those of certain parabolic cylinder type sets. The intersection percentage functions of the latter sets can be described as threefold integrals. It turns out that these intersection percentage functions yield simultaneously geometric representation formulae for the doubly noncentral g-generalized F-distributions. Hence, we get beyond new formulae for evaluating probabilities of correct classification new geometric representation formulae for the doubly noncentral g-generalized F-distributions. A numerical study concerning several aspects of evaluating both probabilities of correct classification and values of the doubly noncentral g-generalized F-distributions demonstrates the advantageous computational properties of the present new approach. This impression will be supported by comparison with the literature.It is shown that probabilities of correct classification depend on the parameters of the underlying sample distribution through a certain well-defined set of secondary parameters. If the underlying parameters are unknown, we propose to estimate probabilities of correct classification.     

Exchangeably weighted bootstraps of empirical estimators of a semi-Markov kernel

A general notion of bootstrapped empirical estimators, of the semi-Markov kernels and of the conditional transition probabilities for semi-Markov processes with countable state space, constructed by exchangeably weighting sample, is introduced. Asymptotic properties of these generalized bootstrapped empirical distributions are obtained by means of the martingale approach.     

Optimal feedback control policies for stochastic,distributed-parameter systems

This paper considers optimal feedback control policies for a class of discrete stochastic distributed-parameter systems. The class under consideration has the property that the random variable in the dynamic systems depends only on the time and possesses the Markovian property with stationary transition probabilities. A necessary condition for optimality of a feedback control policy, which has form similar to the Hamiltonian form in the deterministic case, is derived via a dynamic programming approach.