Objective Bayesianism Research Articles

Empirical Bayes Methods, Evidentialism, and the Inferential Roles They Play.

Empirical Bayes-based Methods (EBM) is an increasingly popular form of Objective Bayesianism (OB). It is identified in particular with the statistician Bradley Efron. The main aims of this paper are, first, to describe and illustrate its main features and, second, to locate its role by comparing it with two other statistical paradigms, Subjective Bayesianism (SB) and Evidentialism. EBM's main formal features are illustrated in some detail by schematic examples. The comparison between what Efron calls their underlying "philosophies" is by way of a distinction made between confirmation and evidence. Although this distinction is sometimes made in the statistical literature, it is relatively rare and never to the same point as here. That is, the distinction is invariably spelled out intra- and not inter-paradigmatically solely in terms of one or the other accounts. The distinction made in this paper between confirmation and evidence is illustrated by two well-known statistical paradoxes: the base-rate fallacy and Popper's paradox of ideal evidence. The general conclusion reached is that each of the paradigms has a basic role to play and all are required by an adequate account of statistical inference from a technically informed and fine-grained philosophical perspective.

Making decisions with evidential probability and objective Bayesian calibration inductive logics

Calibration inductive logics are based on accepting estimates of relative frequencies, which are used to generate imprecise probabilities. In turn, these imprecise probabilities are intended to guide beliefs and decisions — a process called “calibration”. Two prominent examples are Henry E. Kyburg's system of Evidential Probability and Jon Williamson's version of Objective Bayesianism. There are many unexplored questions about these logics. How well do they perform in the short-run? Under what circumstances do they do better or worse? What is their performance relative to traditional Bayesianism?In this article, we develop an agent-based model of a classic binomial decision problem, including players based on variations of Evidential Probability and Objective Bayesianism. We compare the performances of these players, including against a benchmark player who uses standard Bayesian inductive logic. We find that the calibrated players can match the performance of the Bayesian player, but only with particular acceptance thresholds and decision rules. Among other points, our discussion raises some challenges for characterising “cautious” reasoning using imprecise probabilities. Thus, we demonstrate a new way of systematically comparing imprecise probability systems, and we conclude that calibration inductive logics are surprisingly promising for making decisions.

A Battle in the Statistics Wars: a simulation-based comparison of Bayesian, Frequentist and Williamsonian methodologies

The debates between Bayesian, frequentist, and other methodologies of statistics have tended to focus on conceptual justifications, sociological arguments, or mathematical proofs of their long run properties. Both Bayesian statistics and frequentist (“classical”) statistics have strong cases on these grounds. In this article, we instead approach the debates in the “Statistics Wars” from a largely unexplored angle: simulations of different methodologies’ performance in the short to medium run. We used Big Data methods to conduct a large number of simulations using a straightforward decision problem based around tossing a coin with unknown bias and then placing bets. In this simulation, we programmed four players, inspired by Bayesian statistics, frequentist statistics, Jon Williamson’s version of Objective Bayesianism, and a player who simply extrapolates from observed frequencies to general frequencies. The last player served a benchmark function: any worthwhile statistical methodology should at least match the performance of simplistic induction. We focused on the performance of these methodologies in guiding the players towards good decisions. Unlike an earlier simulation study of this type, we found no systematic difference in performance between the Bayesian and frequentist players, provided the Bayesian used a flat prior and the frequentist used a low confidence level. The Williamsonian player was also able to perform well given a low confidence level. However, the frequentist and Williamsonian players performed poorly with high confidence levels, while the Bayesian was surprisingly harmed by biased priors. Our study indicates that all three methodologies should be taken seriously by philosophers and practitioners of statistics.

Parametric boostrap and objective Bayesian testing for heteroscedastic one-way ANOVA

This article explores relationship between parametric bootstrap (PB) and objective Bayesian (OB) approaches to test equality of the factor level means (overall mean test) of heteroscedastic one-way analysis of variance (ANOVA) problem. We compared overall mean tests based on PB, OB, and OB using posterior predictive distribution approaches by simulation studies. We also proved that the PB and OB overall mean tests for one-way heteroscedastic ANOVA are asymptotically equivalent.

Objective Bayesianism and the Abductivist Response to Scepticism

AbstractAn important line of response to scepticism appeals to the best explanation. But anti-sceptics have not engaged much with work on explanation in the philosophy of science. I plan to investigate whether plausible assumptions about best explanations really do favour anti-scepticism. I will argue that there are ways of constructing sceptical hypotheses in which the assumptions do favour anti-scepticism, but the size of the support for anti-scepticism is small.

E.T. Jaynes’s Solution to the Problem of Countable Additivity

Philosophers cannot agree on whether the rule of Countable Additivity should be an axiom of probability. Edwin T. Jaynes attacks the problem in a way which is original to him and passed over in the current debate about the principle: he says the debate only arises because of an erroneous use of mathematical infinity. I argue that this solution fails, but I construct a different argument which, I argue, salvages the spirit of the more general point Jaynes makes. I argue that in Jaynes’s objective Bayesianism we might have good reasons to adopt Countable Additivity, and some of the major problems this adoption is known to entail need not worry us. In particular, I propose to adopt this new angle on Countable Additivity in Jon Williamson’s version of objective Bayesianism.

The Principal Principle and subjective Bayesianism

This paper poses a problem for Lewis’ Principal Principle in a subjective Bayesian framework: we show that, where chances inform degrees of belief, subjective Bayesianism fails to validate normal informal standards of what is reasonable. This problem points to a tension between the Principal Principle and the claim that conditional degrees of belief are conditional probabilities. However, one version of objective Bayesianism has a straightforward resolution to this problem, because it avoids this latter claim. The problem, then, offers some support to this version of objective Bayesianism.

Estimation of type I censored exponential distribution parameters using objective bayesian and bootstrap methods (case study of chronic kidney failure patients)

Bayesian point estimation is an estimation method based on prior selection and loss function. In Objective Bayesian estimation are chosen prior to Jeffrey and used intrinsic discrepancy loss functions based on the Kullback-Leibler divergence equation which will have a minimum effect of data on the posterior distribution. The objective Bayesian point estimator provides estimates of population parameters based solely on the assumed population distribution and data. The goal of this paper is to estimate parameters from the exponential distribution on type II censored data using the objective Bayesian and bootstrap methods. The bootstrap method is used to resampling and built a confidence intervals for parameters whhich will be estimated. The methods were applied on the life-time data of 63 patients of chronic renal failure and the initial diagnosis was non-diabetic disease with bootstrap methods using 10, 100, and 1000 times used in this study. So that the bigger bootstrap samples rendered the estimated value θ̂ will be better and the result confidence interval ranges narrower.

Naturalness as a Constraint on Priors

AbstractMany epistemological problems can be solved by the objective Bayesian view that there are rationality constraints on priors, that is, inductive probabilities. But attempts to work out these constraints have run into such serious problems that many have rejected objective Bayesianism altogether. I argue that the epistemologist should borrow the metaphysician’s concept of naturalness and assign higher priors to more natural hypotheses.

Probabilistic Knowledge and Cognitive Ability

Sarah Moss (2013) argues that degrees of belief, or credences, can amount to knowledge in much the way that full beliefs can. This essay explores a new kind of objective Bayesianism designed to take us some way toward securing such knowledge-constituting credences, or “probabilistic knowledge.” Whatever else it takes for an agent's credences to amount to knowledge, their success, or accuracy, must be the product of cognitive ability or skill. The brand of Bayesianism developed here helps ensure this ability condition is satisfied. Cognitive ability, in turn, helps make credences valuable in other ways: it helps mitigate their dependence on epistemic luck, for example. What we end up with, at the end of the day, are credences that are particularly good candidates for constituting probabilistic knowledge. What's more, examining the character of these credences teaches us something important about what the pursuit of probabilistic knowledge demands from us. It does not demand that we give hypotheses equal treatment, by affording them equal credence. Rather, it demands that we give them equal consideration, by affording them an equal chance of being discovered.

Logical Probability and the Strength of Mathematical Conjectures

Mathematicians often speak of the evidence for unproved conjectures, such as the Riemann Hypothesis. It is argued that such evidence should be seen in terms of logical probability in Keynes's sense: a strictly logical degree of partial implication. That is essentially the same as objective Bayesianism. Examples are given and explained in terms of the objective logical strength of evidence.

Updating a progic

This paper presents a progic, or probabilistic logic, in the sense of Haenni et al. [8]. The progic presented here is based on Bayesianism, as the progic discussed by Williamson [15]. However, the underlying generalised Bayesianism differs from the objective Bayesianism used by Williamson, in the calibration norm, and the liberalisation and interpretation of the reference probability in the norm of equivocation. As a consequence, the updating dynamics of both Bayesianisms differ essentially. Whereas objective Bayesianism is based on a probabilistic re-evaluation, orthodox Bayesianism is based on a probabilistic revision. I formulate a generalised and iterable orthodox Bayesian revision dynamics. This allows to define an updating procedure for the generalised Bayesian progic. The paper compares the generalised Bayesian progic and Williamson's objective Bayesian progic in strength, update dynamics and with respect to language (in)sensitivity.

The Principal Principle Implies the Principle of Indifference

We argue that David Lewis’s principal principle implies a version of the principle of indifference. The same is true for similar principles that need to appeal to the concept of admissibility. Such principles are thus in accord with objective Bayesianism, but in tension with subjective Bayesianism.

Justifying Objective Bayesianism on Predicate Languages

Objective Bayesianism says that the strengths of one’s beliefs ought to be probabilities, calibrated to physical probabilities insofar as one has evidence of them, and otherwise sufficiently equivocal. These norms of belief are often explicated using the maximum entropy principle. In this paper we investigate the extent to which one can provide a unified justification of the objective Bayesian norms in the case in which the background language is a first-order predicate language, with a view to applying the resulting formalism to inductive logic. We show that the maximum entropy principle can be motivated largely in terms of minimising worst-case expected loss.

OBJECTIVE BAYESIAN ANALYSIS OF “ON/OFF” MEASUREMENTS

In high-energy astrophysics, it is common practice to account for the background overlaid with the counts from the source of interest with the help of auxiliary measurements carried on by pointing off-source. In this "on/off" measurement, one knows the number of photons detected while pointing to the source, the number of photons collected while pointing away of the source, and how to estimate the background counts in the source region from the flux observed in the auxiliary measurements. For very faint sources, the number of detected photons is so low that the approximations which hold asymptotically are not valid. On the other hand, the analytical solution exists for the Bayesian statistical inference, which is valid at low and high counts. Here we illustrate the objective Bayesian solution based on the reference posterior and compare the result with the approach very recently proposed by \citet{knoetig2014}, discussing its most delicate points. In addition, we propose to compute the significance of the excess with respect to the background-only expectation with a method which is able to account for any uncertainty on the background and is valid for any photon count. This method is compared to the widely used significance formula by \citet{LiMa83}, which is based on asymptotic properties.

Equivocation for the Objective Bayesian

According to Williamson (In defense of objective Bayesianism, Oxford University Press, Oxford, 2010), the difference between empirical subjective Bayesians and objective Bayesians is that, while both hold reasonable credence to be calibrated to evidence, the objectivist also takes such credence to be as equivocal as such calibration allows. However, Williamson’s prescription for equivocation generates constraints on reasonable credence that are objectionable. Herein Williamson’s calibration norm is explicated in a novel way that permits an alternative equivocation norm. On this alternative account, evidence calibrated probability functions are recognised as implications of evidence calibrated density functions defined over chance hypotheses. The objective Bayesian equivocates between these calibrated density functions rather than between the calibrated probability functions themselves. The result is an objective Bayesianism that avoids the main problem afflicting Williamson’s original proposal.

Objective Bayesianism and the Maximum Entropy Principle

Objective Bayesian epistemology invokes three norms: the strengths of our beliefs should be probabilities; they should be calibrated to our evidence of physical probabilities; and they should otherwise equivocate sufficiently between the basic propositions that we can express. The three norms are sometimes explicated by appealing to the maximum entropy principle, which says that a belief function should be a probability function, from all those that are calibrated to evidence, that has maximum entropy. However, the three norms of objective Bayesianism are usually justified in different ways. In this paper, we show that the three norms can all be subsumed under a single justification in terms of minimising worst-case expected loss. This, in turn, is equivalent to maximising a generalised notion of entropy. We suggest that requiring language invariance, in addition to minimising worst-case expected loss, motivates maximisation of standard entropy as opposed to maximisation of other instances of generalised entropy. Our argument also provides a qualified justification for updating degrees of belief by Bayesian conditionalisation. However, conditional probabilities play a less central part in the objective Bayesian account than they do under the subjective view of Bayesianism, leading to a reduced role for Bayes’ Theorem.

How Uncertain Do We Need to Be?

Expert probability forecasts can be useful for decision making (Sect. 1). But levels of uncertainty escalate: however the forecaster expresses the uncertainty that attaches to a forecast, there are good reasons for her to express a further level of uncertainty, in the shape of either imprecision or higher order uncertainty (Sect. 2). Bayesian epistemology provides the means to halt this escalator, by tying expressions of uncertainty to the propositions expressible in an agent’s language (Sect. 3). But Bayesian epistemology comes in three main varieties. Strictly subjective Bayesianism and empirically-based subjective Bayesianism have difficulty in justifying the use of a forecaster’s probabilities for decision making (Sect. 4). On the other hand, objective Bayesianism can justify the use of these probabilities, at least when the probabilities are consistent with the agent’s evidence (Sect. 5). Hence objective Bayesianism offers the most promise overall for explaining how testimony of uncertainty can be useful for decision making. Interestingly, the objective Bayesian analysis provided in Sect. 5 can also be used to justify a version of the Principle of Reflection (Sect. 6).

Jon Williamson: In Defence of Objective Bayesianism

Jon Williamson: In Defence of Objective Bayesianism

In Defence of Objective Bayesianism

In Defence of Objective Bayesianism