On probability-based inference under data missing by design

Whether a statistician wants to complement a probability model for observed data with a prior distribution and carry out fully probabilistic inference, or base the inference only on the likelihood function, may be a fundamental question in theory, but in practice it may well be of less importance if the likelihood contains much more information than the prior. Maximum likelihood inference can be justified as a Gaussian approximation at the posterior mode, using flat priors. However, in situations where parametric assumptions in standard statistical models would be too rigid, more flexible model formulation, combined with fully probabilistic inference, can be achieved using hierarchical Bayesian parametrization. This work includes five articles, all of which apply probability modeling under various problems involving incomplete observation. Three of the papers apply maximum likelihood estimation and two of them hierarchical Bayesian modeling. Because maximum likelihood may be presented as a special case of Bayesian inference, but not the other way round, in the introductory part of this work we present a framework for probability-based inference using only Bayesian concepts. We also re-derive some results presented in the original articles using the toolbox equipped herein, to show that they are also justifiable under this more general framework. Here the assumption of exchangeability and de Finetti's representation theorem are applied repeatedly for justifying the use of standard parametric probability models with conditionally independent likelihood contributions. It is argued that this same reasoning can be applied also under sampling from a finite population. The main emphasis here is in probability-based inference under incomplete observation due to study design. This is illustrated using a generic two-phase cohort sampling design as an example. The alternative approaches presented for analysis of such a design are full likelihood, which utilizes all observed information, and conditional likelihood, which is restricted to a completely observed set, conditioning on the rule that generated that set. Conditional likelihood inference is also applied for a joint analysis of prevalence and incidence data, a situation subject to both left censoring and left truncation. Other topics covered are model uncertainty and causal inference using posterior predictive distributions. We formulate a non-parametric monotonic regression model for one or more covariates and a Bayesian estimation procedure, and apply the model in the context of optimal sequential treatment regimes, demonstrating that inference based on posterior predictive distributions is feasible also in this case.

Computing the Stochastic Complexity of Simple Probabilistic Graphical Models

Minimum Description Length (MDL) is an information-theoretic principle that can be used for model selection and other statistical inference tasks. There are various ways to use the principle in practice. One theoretically valid way is to use the normalized maximum likelihood (NML) criterion. Due to computational difficulties, this approach has not been used very often. This thesis presents efficient floating-point algorithms that make it possible to compute the NML for multinomial, Naive Bayes and Bayesian forest models. None of the presented algorithms rely on asymptotic analysis and with the first two model classes we also discuss how to compute exact rational number solutions.

A Probabilistic Approach to the Primary Visual Cortex

What can the statistical structure of natural images teach us about the human brain? Even though the visual cortex is one of the most studied parts of the brain, surprisingly little is known about how exactly images are processed to leave us with a coherent percept of the world around us, so we can recognize a friend or drive on a crowded street without any effort. By constructing probabilistic models of natural images, the goal of this thesis is to understand the structure of the stimulus that is the raison d etre for the visual system. Following the hypothesis that the optimal processing has to be matched to the structure of that stimulus, we attempt to derive computational principles, features that the visual system should compute, and properties that cells in the visual system should have. Starting from machine learning techniques such as principal component analysis and independent component analysis we construct a variety of sta- tistical models to discover structure in natural images that can be linked to receptive field properties of neurons in primary visual cortex such as simple and complex cells. We show that by representing images with phase invariant, complex cell-like units, a better statistical description of the vi- sual environment is obtained than with linear simple cell units, and that complex cell pooling can be learned by estimating both layers of a two-layer model of natural images. We investigate how a simplified model of the processing in the retina, where adaptation and contrast normalization take place, is connected to the nat- ural stimulus statistics. Analyzing the effect that retinal gain control has on later cortical processing, we propose a novel method to perform gain control in a data-driven way. Finally we show how models like those pre- sented here can be extended to capture whole visual scenes rather than just small image patches. By using a Markov random field approach we can model images of arbitrary size, while still being able to estimate the model parameters from the data.

A probabilistic approach to the problem of electron localization in disordered systems and sharpness of the mobility edge

By applying the theory of the asymptotic distribution of extremes and a certain stability criterion to the question of the domain of convergence in the probability sense, of the renormalized perturbation expansion (RPE) for the site self-energy in a cellularly disordered system, an expression has been obtained in closed form for the probability of nonconvergence of the RPE on the real-energy axis. Hence, the intrinsic mobility mu (E) as a function of the carrier energy E is deduced to be given by mu (E)= mu 0exp(-exp( mod E mod -Ec) Delta ), where Ec is a nominal 'mobility edge' and Delta is the width of the random site-energy distribution. Thus mobility falls off sharply but continuously for mod E mod >Ec, in contradistinction with the notion of an abrupt 'mobility edge' proposed by Cohen et al. and Mott. Also, the calculated electrical conductivity shows a temperature dependence in qualitative agreement with experiments on disordered semiconductors.

Efficient probabilistic contingency analysis through a stability measure considering wind perturbation

The integration of stochastic wind power has accentuated a challenge for power system stability assessment. Since the power system is a time-variant system under wind generation fluctuations, pure time-domain simulations are difficult to provide real-time stability assessment. As a result, the worst-case scenario is simulated to give a very conservative assessment of system transient stability. In this study, a probabilistic contingency analysis through a stability measure method is proposed to provide a less conservative contingency analysis which covers 5-min wind fluctuations and a successive fault. This probabilistic approach would estimate the transfer limit of a critical line for a given fault with stochastic wind generation and active control devices in a multi-machine system. This approach achieves a lower computation cost and improved accuracy using a new stability measure and polynomial interpolation, and is feasible for online contingency analysis.

Development of probabilistic understanding in fourth grade

We analyzed the development of 4th-grade students’ understanding of the transition from experimental relative frequencies of outcomes to theoretical probabilities with a focus on the foundational statistical concepts of variation and expectation. We report students’ initial and changing expectations of the outcomes of tossing one and two coins, how they related the relative frequency from their physical and computersimulated trials to the theoretical probability, and how they created and interpreted theoretical probability models. Findings include students’ progression from an initial apparent equiprobability bias in predicting outcomes of tossing two coins through to representing the outcomes of increasing the number of trials. After observing the decreasing variation from the theoretical probability as the sample size increased, students developed a deeper understanding of the relationship between relative frequency of outcomes and theoretical probability as well as their respective associations with variation and expectation. Students’ final models indicated increasing levels of probabilistic understanding.

Probabilistic Inductive Querying Using ProbLog

We study how probabilistic reasoning and inductive querying can be combined within ProbLog, a recent probabilistic extension of Prolog. ProbLog can be regarded as a database system that supports both probabilistic and inductive reasoning through a variety of querying mechanisms. After a short introduction to ProbLog, we provide a survey of the different types of inductive queries that ProbLog supports, and show how it can be applied to the mining of large biological networks.

Open Access to the Scientific Journal Literature: Situation 2009

Background: The Internet has recently made possible the free global availability of scientific journal articles. Open Access (OA) can occur either via OA scientific journals, or via authors posting manuscripts of articles published in subscription journals in open web repositories. So far there have been few systematic studies showing how big the extent of OA is, in particular studies covering all fields of science. Methodology/Principal Findings: The proportion of peer reviewed scholarly journal articles, which are available openly in full text on the web, was studied using a random sample of 1837 titles and a web search engine. Of articles published in 2008, 8,5% were freely available at the publishers’ sites. For an additional 11,9% free manuscript versions could be found using search engines, making the overall OA percentage 20,4%. Chemistry (13%) had the lowest overall share of OA, Earth Sciences (33%) the highest. In medicine, biochemistry and chemistry publishing in OA journals was more common. In all other fields author-posted manuscript copies dominated the picture. Conclusions/Significance: The results show that OA already has a significant positive impact on the availability of the scientific journal literature and that there are big differences between scientific disciplines in the uptake. Due to the lack of awareness of OA-publishing among scientists in most fields outside physics, the results should be of general interest to all scholars. The results should also interest academic publishers, who need to take into account OA in their business strategies and copyright policies, as well as research funders, who like the NIH are starting to require OA availability of results from research projects they fund. The method and search tools developed also offer a good basis for more in-depth studies as well as longitudinal studies.

Probabilistic evaluation of seismic soil liquefaction potential based on SPT data

The performance-based liquefaction potential analysis was carried out in the present study to estimate the liquefaction return period for Bangalore, India, through a probabilistic approach. In this approach, the entire range of peak ground acceleration (PGA) and earthquake magnitudes was used in the evaluation of liquefaction return period. The seismic hazard analysis for the study area was done using probabilistic approach to evaluate the peak horizontal acceleration at bed rock level. Based on the results of the multichannel analysis of surface wave, it was found that the study area belonged to site class D. The PGA values for the study area were evaluated for site class D by considering the local site effects. The soil resistance for the study area was characterized using the standard penetration test (SPT) values obtained from 450 boreholes. These SPT data along with the PGA values obtained from the probabilistic seismic hazard analysis were used to evaluate the liquefaction return period for the study area. The contour plot showing the spatial variation of factor of safety against liquefaction and the corrected SPT values required for preventing liquefaction for a return period of 475 years at depths of 3 and 6 m are presented in this paper. The entire process of liquefaction potential evaluation, starting from collection of earthquake data, identifying the seismic sources, evaluation of seismic hazard and the assessment of liquefaction return period were carried out, and the entire analysis was done based on the probabilistic approach.

Probabilistic Analysis of Cracking Moment of Reinforced Concrete Beams

Probabilistic analysis of cracking moment from 22 simply supported reinforced concrete beams is performed. When the basic variables follow the distribution considered in this study, the cracking moment of a beam is found to follow a normal distribution. An expression is derived, for characteristic cracking moment, which will be useful in examining reinforced concrete beams for a limit state of cracking.

Solid-state chemistry of high-temperature oxide superconductors:The experimental situation

High-temperature superconductivity in oxides of the type(La, Ln)2?xBax(Sr)xCuO4, Y(Ln)Ba2Cu3O7??, La3?xBa3+xCu6O14, and related systems is discussed with emphasis on aspects related to experimental solid-state chemistry. All of these oxides possess perovskite-related structures. Oxygen-excess and La-deficient La2CuO4 also exhibit superconductivity in the 20�40 K just as La2?xBax(Srx)CuO4; these oxides are orthorhombic in the superconductivity phase. The crucial role of oxygen stoichiometry in the superconductivity ofYBa2Cu3O7?? (Tc = 95 ± 5K) is examined; this oxide remains orthorhombic up to ? ? 0.6 and becomes tetragonal and nonsuperconducting beyond this value of ?. Oxygen stoichiometry in this and related oxides has to be understood in terms of structure and disorder. The structure of La3?xBa3+xCu6O14 is related to that of YBa2Cu3O7, the orthorhombic structure manifesting itself when the population of O1 oxygens (along the Cusingle bondOsingle bondCu chains) is preponderant compared to that of O5 oxygens (along thea-axis); nearly equal populations of O1 and O5 sites give rise to the tetragonal structure. A transition from a high-Tc (95 K) superconductivity regime to a low-Tc (not, vert, similar60 K) regime occurs in YBa2Cu3O7?? accompanying a change in ?. There is no evidence for Cu3+ in these nominally mixed valent copper oxides. Instead, holes are present on oxygens giving rise to O? or O2?2 species, the concentration of these species increasing with the lowering of temperature. Certain interesting aspects of the superconducting oxides such as domain or twin boundaries, Raman spectra, microwave absorption, and anomalous high-temperature resistivity drops are presented along with the important material parameters. Preparative aspects of the superconducting oxides are briefly discussed. Phase transitions seem to occur atTc as well as at not, vert, similar240 K in YBa2Cu3O7.

Probabilistic extension to realistic abductive reasoning model

In this paper, we give a method for probabilistic assignment to the Realistic Abductive Reasoning Model, The knowledge is assumed to be represented in the form of causal chaining, namely, hyper-bipartite network. Hyper-bipartite network is the most generalized form of knowledge representation for which, so far, there has been no way of assigning probability to the explanations, First, the inference mechanism using realistic abductive reasoning model is briefly described and then probability is assigned to each of the explanations so as to pick up the explanations in the decreasing order of plausibility.

A checking method for probabilistic seismic-hazard assessment: case studies on three cities

The conventional Cornell's source-based approach of probabilistic seismic-hazard assessment (PSHA) has been employed all around the world, whilst many studies often rely on the use of computer packages such as FRISK (McGuire FRISK-a computer program for seismic risk analysis. Open-File Report 78-1007, United States Geological Survey, Department of Interior, Washington 1978) and SEISRISK III (Bender and Perkins SEISRISK III-a computer program for seismic hazard estimation, Bulletin 1772. United States Geological Survey, Department of Interior, Washington 1987). A ``black-box'' syndrome may be resulted if the user of the software does not have another simple and robust PSHA method that can be used to make comparisons. An alternative method for PSHA, namely direct amplitude-based (DAB) approach, has been developed as a heuristic and efficient method enabling users to undertake their own sanity checks on outputs from computer packages. This paper experiments the application of the DAB approach for three cities in China, Iran, and India, respectively, and compares with documented results computed by the source-based approach. Several insights regarding the procedure of conducting PSHA have also been obtained, which could be useful for future seismic-hazard studies.