Bayesian graphical model determination using decision theory

Bayesian model determination in the complete class of graphical models is considered using a decision theoretic framework within the regular exponential family. The complete class contains both decomposable and non-decomposable graphical models. A utility measure based on a logarithmic score function is introduced under reference priors for the model parameters. The logarithmic utility of a model is decomposed into predictive performance and relative complexity. Axioms of decision theory lead to the judgement of the plausibility of a model in terms of the posterior expected utility. This quantity has an analytic expression for decomposable models when certain reference priors are used and the exponential family is closed under marginalization. For non-decomposable models, a simulation consistent estimate of the expectation can be obtained. Both real and simulated data sets are used to illustrate the introduced methodology.     

Properties of graphical regression models for multidimensional categorical data

We propose a two-component graphical chain model, the discrete regression distribution, where a set of discrete random variables is modeled as a response to a set of categorical and continuous covariates. The proposed model is useful for modeling a set of discrete variables measured at multiple sites along with a set of continuous and/or discrete covariates. The proposed model allows for joint examination of the dependence structure of the discrete response and observed covariates and also accommodates site-to-site variability. We develop the graphical model properties and theoretical justifications of this model. Our model has several advantages over the traditional logistic normal model used to analyze similar compositional data, including site-specific random effect terms and the incorporation of discrete and continuous covariates.     

Approaches to multistage one-shot decision making

In this research, multistage one-shot decision making under uncertainty is studied. In such a decision problem, a decision maker has one and only one chance to make a decision at each stage with possibilistic information. Based on the one-shot decision theory, approaches to multistage one-shot decision making are proposed. In the proposed approach, a decision maker chooses one state amongst all the states according to his/her attitude about satisfaction and possibility at each stage. The payoff at each stage is associated with the focus points at the succeeding stages. Based on the selected states (focus points), the sequence of optimal decisions is determined by dynamic programming. The proposed method is a fundamental alternative for multistage decision making under uncertainty because it is scenario-based instead of lottery-based as in the other existing methods. The one-shot optimal stopping problem is analyzed where a decision maker has only one chance to determine stopping or continuing at each stage. The theoretical results have been obtained.     

Approximate credal network updating by linear programming with applications to decision making

Credal nets are probabilistic graphical models which extend Bayesian nets to cope with sets of distributions. An algorithm for approximate credal network updating is presented. The problem in its general formulation is a multilinear optimization task, which can be linearized by an appropriate rule for fixing all the local models apart from those of a single variable. This simple idea can be iterated and quickly leads to accurate inferences. A transformation is also derived to reduce decision making in credal networks based on the maximality criterion to updating. The decision task is proved to have the same complexity of standard inference, being NP^PP-complete for general credal nets and NP-complete for polytrees. Similar results are derived for the E-admissibility criterion. Numerical experiments confirm a good performance of the method.     

Two optimal strategies for active learning of causal models from interventional data

From observational data alone, a causal DAG is only identifiable up to Markov equivalence. Interventional data generally improves identifiability; however, the gain of an intervention strongly depends on the intervention target, that is, the intervened variables. We present active learning (that is, optimal experimental design) strategies calculating optimal interventions for two different learning goals. The first one is a greedy approach using single-vertex interventions that maximizes the number of edges that can be oriented after each intervention. The second one yields in polynomial time a minimum set of targets of arbitrary size that guarantees full identifiability. This second approach proves a conjecture of Eberhardt (2008) [1] indicating the number of unbounded intervention targets which is sufficient and in the worst case necessary for full identifiability. In a simulation study, we compare our two active learning approaches to random interventions and an existing approach, and analyze the influence of estimation errors on the overall performance of active learning.     

An Efficient Procedure for Dynamic Lot-sizing Model with Demand Time Windows

We consider a dynamic lot-sizing model with demand time windows where n demands need to be scheduled in T production periods. For the case of backlogging allowed, an O(T ³) algorithm exists under the non-speculative cost structure. For the same model with somewhat general cost structure, we propose an efficient algorithm with O(max {T ², nT}) time complexity.     

An interactive procedure for linguistic multiple attribute decision making with incomplete weight information

In this paper, we consider the multiple attribute decision making (MADM) problems, in which the information about attribute weights is partly known and the attribute values are expressed in linguistic labels. We first define the concepts of linguistic positive ideal point, linguistic negative ideal point, and satisfactory degree of alternative. Based on these concepts, we then establish some linear programming models, through which the decision maker interacts with the analyst. Furthermore, we establish a practical interactive procedure for solving the MADM problems considered in this paper. The interactive process can be realized by giving and revising the satisfactory degrees of alternatives till an optimum satisfactory solution is achieved. Finally, a practical example is given to illustrate the developed procedure.     

A pseudo-Bayesian model in financial decision making with implications to market volatility,under- and overreaction

This paper develops a model of weight assignments using a pseudo-Bayesian approach that reflects investors’ behavioral biases. In this parsimonious model of investor sentiment, weights induced by investors’ conservative and representative heuristics are assigned to observations of the earning shocks of stock prices. Such weight assignments enable us to provide a quantitative link between some market anomalies and investors’ behavioral biases. The seriousness of an anomaly can be quantitatively assessed by investigating into its dependency on weights. New results other than the short-run underreaction and long-run overreaction can be derived and new hypotheses can be formed.     

Parallel computing applied to the stochastic dynamic programming for long term operation planning of hydrothermal power systems

In this paper, parallel processing techniques are employed to improve the performance of the stochastic dynamic programming applied to the long term operation planning of electrical power system. The hydroelectric plants are grouped into energy equivalent reservoirs and the expected cost functions are modeled by a piecewise linear approximation, by means of the Convex Hull algorithm. In order to validate the proposed methodology, data from the Brazilian electrical power system is utilized.