Adaptively weighted maximum likelihood estimation of discrete distributions

SummaryDiscrete data arise in various research fields, typically when the observations are count data.I propose a robust and efficient parametric procedure for estimation of discrete distributions. The estimation is done in two phases. First, a very robust, but possibly inefficient, estimate of the model parameters is computed and used to indentify outliers. Then the outliers are either removed from the sample or given low weights, and a weighted maximum likelihood estimate (WML) is computed.The weights are determined via an adaptive process such that if the data follow the model, then asymptotically no observation is downweighted.I prove that the final estimator inherits the breakdown point of the initial one, and that its influence function at the model is the same as the influence function of the maximum likelihood estimator, which strongly suggests that it is asymptotically fully efficient.The initial estimator is a minimum disparity estimator (MDE). MDEs can be shown to have full asymptotic efficiency, and some MDEs have very high breakdown points and very low bias under contamination. Several initial estimators are considered, and the performances of the WMLs based on each of them are studied.It results that in a great variety of situations the WML substantially improves the initial estimator, both in terms of finite sample mean square error and in terms of bias under contamination. Besides, the performances of the WML are rather stable under a change of the MDE even if the MDEs have very different behaviors.Two examples of application of the WML to real data are considered. In both of them, the necessity for a robust estimator is clear: the maximum likelihood estimator is badly corrupted by the presence of a few outliers.This procedure is particularly natural in the discrete distribution setting, but could be extended to the continuous case, for which a possible procedure is sketched.RésuméLes données discrètes sont présentes dans différents domaines de recherche, en particulier lorsque les observations sont des comptages.Je propose une méthode paramétrique robuste et efficace pour l'estimation de distributions discrètes. L'estimation est faite en deux phases. Tout d'abord, un estimateur très robuste des paramètres du modèle est calculé, et utilisé pour la détection des données aberrantes (outliers). Cet estimateur n'est pas nécessairement efficace. Ensuite, soit les outliers sont retirés de l'échantillon, soit des faibles poids leur sont attribués, et un estimateur du maximum de vraisemblance pondéré (WML) est calculé.Les poids sont déterminés via un processus adaptif, tel qu'asymptotiquement, si les données suivent le modèle, aucune observation n'est dépondérée.Je prouve que le point de rupture de l'estimateur final est au moins aussi élevé que celui de l'estimateur initial, et que sa fonction d'influence au modèle est la même que celle du maximum de vraisemblance, ce qui suggère que cet estimateur est pleinement efficace asymptotiquement.L'estimateur initial est un estimateur de disparité minimale (MDE). Les MDE sont asymptotiquement pleinement efficaces, et certains d'entre eux ont un point de rupture très élevé et un très faible biais sous contamination. J'étudie les performances du WML basé sur différents MDEs.Le résultat est que dans une grande variété de situations le WML améliore largement les performances de l'estimateur initial, autant en terme du carré moyen de l'erreur que du biais sous contamination. De plus, les performances du WML restent assez stables lorsqu'on change l'estimateur initial, même si les différents MDEs ont des comportements très différents.Je considère deux exemples d'application du WML à des données réelles, où la nécessité d'un estimateur robuste est manifeste : l'estimateur du maximum de vraisemblance est fortement corrompu par la présence de quelques outliers.La méthode proposée est particulièrement naturelle dans le cadre des distributions discrètes, mais pourrait être étendue au cas continu.

Ab initio absorption spectrum of NiO combining molecular dynamics with the embedded cluster approach in a discrete reaction field

We developed a procedure that combines three complementary computational methodologies to improve the theoretical description of the electronic structure of nickel oxide. The starting point is a Car-Parrinello molecular dynamics simulation to incorporate vibrorotational degrees of freedom into the material model. By means ofcomplete active space self-consistent field second-order perturbation theory (CASPT2) calculations on embedded clusters extracted from the resulting trajectory, we describe localized spectroscopic phenomena on NiO with an efficient treatment of electron correlation. The inclusion of thermal motion into the theoretical description allowsus to study electronic transitions that, otherwise, would be dipole forbidden in the ideal structure and results in a natural reproduction of the band broadening. Moreover, we improved the embedded cluster model by incorporating self-consistently at the complete active space self-consistent field (CASSCF) level a discrete (or direct) reaction field (DRF) in the cluster surroundings. The DRF approach offers an efficient treatment ofelectric response effects of the crystalline embedding to the electronic transitions localized in the cluster. We offer accurate theoretical estimates of the absorption spectrum and the density of states around the Fermi level of NiO, and a comprehensive explanation of the source of the broadening and the relaxation of the charge transferstates due to the adaptation of the environment

Body fluid and tissue analysis using filter paper sampling support prior to LC-MS/MS: application to fatal overdose with colchicine

Because of the various matrices available for forensic investigations, the development of versatile analytical approaches allowing the simultaneous determination of drugs is challenging. The aim of this work was to assess a liquid chromatography-tandem mass spectrometry (LC-MS/MS) platform allowing the rapid quantification of colchicine in body fluids and tissues collected in the context of a fatal overdose. For this purpose, filter paper was used as a sampling support and was associated with an automated 96-well plate extraction performed by the LC autosampler itself. The developed method features a 7-min total run time including automated filter paper extraction (2 min) and chromatographic separation (5 min). The sample preparation was reduced to a minimum regardless of the matrix analyzed. This platform was fully validated for dried blood spots (DBS) in the toxic concentration range of colchicine. The DBS calibration curve was applied successfully to quantification in all other matrices (body fluids and tissues) except for bile, where an excessive matrix effect was found. The distribution of colchicine for a fatal overdose case was reported as follows: peripheral blood, 29 ng/ml; urine, 94 ng/ml; vitreous humour and cerebrospinal fluid, < 5 ng/ml; pericardial fluid, 14 ng/ml; brain, < 5 pg/mg; heart, 121 pg/mg; kidney, 245 pg/mg; and liver, 143 pg/mg. Although filter paper is usually employed for DBS, we report here the extension of this alternative sampling support to the analysis of other body fluids and tissues. The developed platform represents a rapid and versatile approach for drug determination in multiple forensic media.

Wipe sampling procedure coupled to LC-MS/MS analysis for the simultaneous determination of 10 cytotoxic drugs on different surfaces.

A simple wipe sampling procedure was developed for the surface contamination determination of ten cytotoxic drugs: cytarabine, gemcitabine, methotrexate, etoposide phosphate, cyclophosphamide, ifosfamide, irinotecan, doxorubicin, epirubicin and vincristine. Wiping was performed using Whatman filter paper on different surfaces such as stainless steel, polypropylene, polystyrol, glass, latex gloves, computer mouse and coated paperboard. Wiping and desorption procedures were investigated: The same solution containing 20% acetonitrile and 0.1% formic acid in water gave the best results. After ultrasonic desorption and then centrifugation, samples were analysed by a validated liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS) in selected reaction monitoring mode. The whole analytical strategy from wipe sampling to LC-MS/MS analysis was evaluated to determine quantitative performance. The lowest limit of quantification of 10 ng per wiping sample (i.e. 0.1 ng cm(-2)) was determined for the ten investigated cytotoxic drugs. Relative standard deviation for intermediate precision was always inferior to 20%. As recovery was dependent on the tested surface for each drug, a correction factor was determined and applied for real samples. The method was then successfully applied at the cytotoxic production unit of the Geneva University Hospitals pharmacy.

Mutual information of IID complex Gaussian signals on block Rayleigh-faded channels

We present a method to compute, quickly and efficiently, the mutual information achieved by an IID (independent identically distributed) complex Gaussian signal on a block Rayleigh-faded channel without side information at the receiver. The method accommodates both scalar and MIMO (multiple-input multiple-output) settings. Operationally, this mutual information represents the highest spectral efficiency that can be attained using Gaussiancodebooks. Examples are provided that illustrate the loss in spectral efficiency caused by fast fading and how that loss is amplified when multiple transmit antennas are used. These examples are further enriched by comparisons with the channel capacity under perfect channel-state information at the receiver, and with the spectral efficiency attained by pilot-based transmission.

Irregular to regular sampling, denoising and deconvolution

We propose a restoration algorithm for band limited images that considers irregular(perturbed) sampling, denoising, and deconvolution. We explore the application of a family ofregularizers that allow to control the spectral behavior of the solution combined with the irregular toregular sampling algorithms proposed by H.G. Feichtinger, K. Gr¨ochenig, M. Rauth and T. Strohmer.Moreover, the constraints given by the image acquisition model are incorporated as a set of localconstraints. And the analysis of such constraints leads to an early stopping rule meant to improvethe speed of the algorithm. Finally we present experiments focused on the restoration of satellite images, where the micro-vibrations are responsible of the type of distortions we are considering here. We will compare results of the proposed method with previous methods and show an extension tozoom.

Mutual Information of IID complex Gaussian signals on block Rayleigh-faded channels

We present a method to compute, quickly and efficiently, the mutual information achieved by an IID (independent identically distributed) complex Gaussian signal on a block Rayleigh-faded channel without side information at the receiver. The method accommodates both scalar and MIMO (multiple-input multiple-output) settings. Operationally, this mutual information represents the highest spectral efficiency that can be attained using Gaussiancodebooks. Examples are provided that illustrate the loss in spectral efficiency caused by fast fading and how that loss is amplified when multiple transmit antennas are used. These examples are further enriched by comparisons with the channel capacity under perfect channel-state information at the receiver, and with the spectral efficiency attained by pilot-based transmission.

Capacity to within one bit of a class of Gaussian multicast channels with interference

This paper studies the fundamental operational limits of a class of Gaussian multicast channels with an interference setting. In particular, the paper considers two base stations multicasting separate messages to distinct sets of users. In the presence of channel state information at the transmitters and at the respective receivers, the capacity region of the Gaussian multicast channel with interference is characterized to within one bit. At the crux of this result is an extension to the multicast channel with interference of the Han-Kobayashi or the Chong-Motani-Garg achievable region for the interference channel.

An Automatic Approach for Learning and Tuning Gaussian Interval Type-2 Fuzzy Membership Functions Applied to Lung CAD Classification System

The potential of type-2 fuzzy sets for managing high levels of uncertainty in the subjective knowledge of experts or of numerical information has focused on control and pattern classification systems in recent years. One of the main challenges in designing a type-2 fuzzy logic system is how to estimate the parameters of type-2 fuzzy membership function (T2MF) and the Footprint of Uncertainty (FOU) from imperfect and noisy datasets. This paper presents an automatic approach for learning and tuning Gaussian interval type-2 membership functions (IT2MFs) with application to multi-dimensional pattern classification problems. T2MFs and their FOUs are tuned according to the uncertainties in the training dataset by a combination of genetic algorithm (GA) and crossvalidation techniques. In our GA-based approach, the structure of the chromosome has fewer genes than other GA methods and chromosome initialization is more precise. The proposed approach addresses the application of the interval type-2 fuzzy logic system (IT2FLS) for the problem of nodule classification in a lung Computer Aided Detection (CAD) system. The designed IT2FLS is compared with its type-1 fuzzy logic system (T1FLS) counterpart. The results demonstrate that the IT2FLS outperforms the T1FLS by more than 30% in terms of classification accuracy.

Discrete devaluations and multiple equilibria in a first generation model of currency crises

The first generation models of currency crises have often been criticized because they predict that, in the absence of very large triggering shocks, currency attacks should be predictable and lead to small devaluations. This paper shows that these features of first generation models are not robust to the inclusion of private information. In particular, this paper analyzes a generalization of the Krugman-Flood-Garber (KFG) model, which relaxes the assumption that all consumers are perfectly informed about the level of fundamentals. In this environment, the KFG equilibrium of zero devaluation is only one of many possible equilibria. In all the other equilibria, the lack of perfect information delays the attack on the currency past the point at which the shadow exchange rate equals the peg, giving rise to unpredictable and discrete devaluations.

Health care provider choice in the case of patient-initiated contacts. An extended version of discrete choice of model demand

This paper analyzes the nature of health care provider choice inthe case of patient-initiated contacts, with special reference toa National Health Service setting, where monetary prices are zeroand general practitioners act as gatekeepers to publicly financedspecialized care. We focus our attention on the factors that mayexplain the continuously increasing use of hospital emergencyvisits as opposed to other provider alternatives. An extendedversion of a discrete choice model of demand for patient-initiatedcontacts is presented, allowing for individual and town residencesize differences in perceived quality (preferences) betweenalternative providers and including travel and waiting time asnon-monetary costs. Results of a nested multinomial logit model ofprovider choice are presented. Individual choice betweenalternatives considers, in a repeated nested structure, self-care,primary care, hospital and clinic emergency services. Welfareimplications and income effects are analyzed by computingcompensating variations, and by simulating the effects of userfees by levels of income. Results indicate that compensatingvariation per visit is higher than the direct marginal cost ofemergency visits, and consequently, emergency visits do not appearas an inefficient alternative even for non-urgent conditions.

Revenue management under general discrete choice model of consumer behavior

Customer choice behavior, such as 'buy-up' and 'buy-down', is an importantphe-nomenon in a wide range of industries. Yet there are few models ormethodologies available to exploit this phenomenon within yield managementsystems. We make some progress on filling this void. Specifically, wedevelop a model of yield management in which the buyers' behavior ismodeled explicitly using a multi-nomial logit model of demand. Thecontrol problem is to decide which subset of fare classes to offer ateach point in time. The set of open fare classes then affects the purchaseprobabilities for each class. We formulate a dynamic program todetermine the optimal control policy and show that it reduces to a dynamicnested allocation policy. Thus, the optimal choice-based policy caneasily be implemented in reservation systems that use nested allocationcontrols. We also develop an estimation procedure for our model based onthe expectation-maximization (EM) method that jointly estimates arrivalrates and choice model parameters when no-purchase outcomes areunobservable. Numerical results show that this combined optimization-estimation approach may significantly improve revenue performancerelative to traditional leg-based models that do not account for choicebehavior.

What risks do people perceive in everyday life? A perspective gained from the experience sampling method (ESM)

The experiential sampling method (ESM) was used to collect data from 74 parttimestudents who described and assessed the risks involved in their current activitieswhen interrupted at random moments by text messages. The major categories ofperceived risk were short-term in nature and involved loss of time or materials relatedto work and physical damage (e.g., from transportation). Using techniques of multilevelanalysis, we demonstrate effects of gender, emotional state, and types of risk onassessments of risk. Specifically, females do not differ from males in assessing thepotential severity of risks but they see these as more likely to occur. Also, participantsassessed risks to be lower when in more positive self-reported emotional states. Wefurther demonstrate the potential of ESM by showing that risk assessments associatedwith current actions exceed those made retrospectively. We conclude by notingadvantages and disadvantages of ESM for collecting data about risk perceptions.