Quantile-Based Optimization of Noisy Computer Experiments with Tunable Precision

This article addresses the issue of kriging-based optimization of stochastic simulators. Many of these simulators depend on factors that tune the level of precision of the response, the gain in accuracy being at a price of computational time. The contribution of this work is two-fold: first, we propose a quantile-based criterion for the sequential design of experiments, in the fashion of the classical expected improvement criterion, which allows an elegant treatment of heterogeneous response precisions. Second, we present a procedure for the allocation of the computational time given to each measurement, allowing a better distribution of the computational effort and increased efficiency. Finally, the optimization method is applied to an original application in nuclear criticality safety. This article has supplementary material available online. The proposed criterion is available in the R package DiceOptim.

Reference values of mechanical and thermal pain tests in a pain-free population

Quantitative sensory tests are widely used in human research to evaluate the effect of analgesics and explore altered pain mechanisms, such as central sensitization. In order to apply these tests in clinical practice, knowledge of reference values is essential. The aim of this study was to determine the reference values of pain thresholds for mechanical and thermal stimuli, as well as withdrawal time for the cold pressor test in 300 pain-free subjects. Pain detection and pain tolerance thresholds to pressure, heat and cold were determined at three body sites: (1) lower back, (2) suprascapular region and (3) second toe (for pressure) or the lateral aspect of the leg (for heat and cold). The influences of gender, age, height, weight, body-mass index (BMI), body side of testing, depression, anxiety, catastrophizing and parameters of Short-Form 36 (SF-36) were analyzed by multiple regressions. Quantile regressions were performed to define the 5th, 10th and 25th percentiles as reference values for pain hypersensitivity and the 75th, 90th and 95th percentiles as reference values for pain hyposensitivity. Gender, age and/or the interaction of age with gender were the only variables that consistently affected the pain measures. Women were more pain sensitive than men. However, the influence of gender decreased with increasing age. In conclusion, normative values of parameters related to pressure, heat and cold pain stimuli were determined. Reference values have to be stratified by body region, gender and age. The determination of these reference values will now allow the clinical application of the tests for detecting abnormal pain reactions in individual patients.

The latency distribution of motor evoked potentials in patients with multiple sclerosis

OBJECTIVE: To compare the individual latency distributions of motor evoked potentials (MEP) in patients with multiple sclerosis (MS) to the previously reported results in healthy subjects (Firmin et al., 2011). METHODS: We applied the previously reported method to measure the distribution of MEP latencies to 16 patients with MS. The method is based on transcranial magnetic stimulation and consists of a combination of the triple stimulation technique with a method originally developed to measure conduction velocity distributions in peripheral nerves. RESULTS: MEP latency distributions in MS typically showed two peaks. The individual MEP latency distributions were significantly wider in patients with MS than in healthy subjects. The mean triple stimulation delay extension at the 75% quantile, a proxy for MEP latency distribution width, was 7.3ms in healthy subjects and 10.7ms in patients with MS. CONCLUSIONS: In patients with MS, slow portions of the central motor pathway contribute more to the MEP than in healthy subjects. The bimodal distribution found in healthy subjects is preserved in MS. SIGNIFICANCE: Our method to measure the distribution of MEP latencies is suitable to detect alterations in the relative contribution of corticospinal tract portions with long MEP latencies to motor conduction.

Characterisation of flash floods in small ungauged mountain basins of Central Spain using an integrated approach

One of the main problems of flood hazard assessment in ungauged or poorly gauged basins is the lack of runoff data. In an attempt to overcome this problem we have combined archival records, dendrogeomorphic time series and instrumental data (daily rainfall and discharge) from four ungauged and poorly gauged mountain basins in Central Spain with the aim of reconstructing and compiling information on 41 flash flood events since the end of the 19th century. Estimation of historical discharge and the incorporation of uncertainty for the at-site and regional flood frequency analysis were performed with an empirical rainfall–runoff assessment as well as stochastic and Bayesian Markov Chain Monte Carlo (MCMC) approaches. Results for each of the ungauged basins include flood frequency, severity, seasonality and triggers (synoptic meteorological situations). The reconstructed data series clearly demonstrates how uncertainty can be reduced by including historical information, but also points to the considerable influence of different approaches on quantile estimation. This uncertainty should be taken into account when these data are used for flood risk management.

Non-exponential return time distributions for vorticity extremes explained by fractional Poisson processes

Serial correlation of extreme midlatitude cyclones observed at the storm track exits is explained by deviations from a Poisson process. To model these deviations, we apply fractional Poisson processes (FPPs) to extreme midlatitude cyclones, which are defined by the 850 hPa relative vorticity of the ERA interim reanalysis during boreal winter (DJF) and summer (JJA) seasons. Extremes are defined by a 99% quantile threshold in the grid-point time series. In general, FPPs are based on long-term memory and lead to non-exponential return time distributions. The return times are described by a Weibull distribution to approximate the Mittag–Leffler function in the FPPs. The Weibull shape parameter yields a dispersion parameter that agrees with results found for midlatitude cyclones. The memory of the FPP, which is determined by detrended fluctuation analysis, provides an independent estimate for the shape parameter. Thus, the analysis exhibits a concise framework of the deviation from Poisson statistics (by a dispersion parameter), non-exponential return times and memory (correlation) on the basis of a single parameter. The results have potential implications for the predictability of extreme cyclones.

Socioeconomic, Temporal and Regional Variation in Body Mass Index among 188,537 Swiss Male Conscripts Born between 1986 and 1992

BACKGROUND Rising levels of overweight and obesity are important public-health concerns worldwide. The purpose of this study is to elucidate their prevalence and trends in Switzerland by analyzing variations in Body Mass Index (BMI) of Swiss conscripts. METHODS The conscription records were provided by the Swiss Army. This study focussed on conscripts 18.5-20.5 years of age from the seven one-year birth cohorts spanning the period 1986-1992. BMI across professional status, area-based socioeconomic position (abSEP), urbanicity and regions was analyzed. Two piecewise quantile regression models with linear splines for three birth-cohort groups were used to examine the association of median BMI with explanatory variables and to determine the extent to which BMI has varied over time. RESULTS The study population consisted of 188,537 individuals. Median BMI was 22.51 kg/m2 (22.45-22.57 95% confidence interval (CI)). BMI was lower among conscripts of high professional status (-0.46 kg/m2; 95% CI: -0.50, -0.42, compared with low), living in areas of high abSEP (-0.11 kg/m2; 95% CI: -0.16, -0.07 compared to medium) and from urban communities (-0.07 kg/m2; 95% CI: -0.11, -0.03, compared with peri-urban). Comparing with Midland, median BMI was highest in the North-West (0.25 kg/m2; 95% CI: 0.19-0.30) and Central regions (0.11 kg/m2; 95% CI: 0.05-0.16) and lowest in the East (-0.19 kg/m2; 95% CI: -0.24, -0.14) and Lake Geneva regions (-0.15 kg/m2; 95% CI: -0.20, -0.09). Trajectories of regional BMI growth varied across birth cohorts, with median BMI remaining high in the Central and North-West regions, whereas stabilization and in some cases a decline were observed elsewhere. CONCLUSIONS BMI of Swiss conscripts is associated with individual and abSEP and urbanicity. Results show regional variation in the levels and temporal trajectories of BMI growth and signal their possible slowdown among recent birth cohorts.

Circulating microRNA profiling in patients with advanced non-squamous NSCLC receiving bevacizumab/erlotinib followed by platinum-based chemotherapy at progression (SAKK 19/05).

OBJECTIVES Molecular subclassification of non small-cell lung cancer (NSCLC) is essential to improve clinical outcome. This study assessed the prognostic and predictive value of circulating micro-RNA (miRNA) in patients with non-squamous NSCLC enrolled in the phase II SAKK (Swiss Group for Clinical Cancer Research) trial 19/05, receiving uniform treatment with first-line bevacizumab and erlotinib followed by platinum-based chemotherapy at progression. MATERIALS AND METHODS Fifty patients with baseline and 24 h blood samples were included from SAKK 19/05. The primary study endpoint was to identify prognostic (overall survival, OS) miRNA's. Patient samples were analyzed with Agilent human miRNA 8x60K microarrays, each glass slide formatted with eight high-definition 60K arrays. Each array contained 40 probes targeting each of the 1347 miRNA. Data preprocessing included quantile normalization using robust multi-array average (RMA) algorithm. Prognostic and predictive miRNA expression profiles were identified by Spearman's rank correlation test (percentage tumor shrinkage) or log-rank testing (for time-to-event endpoints). RESULTS Data preprocessing kept 49 patients and 424 miRNA for further analysis. Ten miRNA's were significantly associated with OS, with hsa-miR-29a being the strongest prognostic marker (HR=6.44, 95%-CI 2.39-17.33). Patients with high has-miR-29a expression had a significantly lower survival at 10 months compared to patients with a low expression (54% versus 83%). Six out of the 10 miRNA's (hsa-miRN-29a, hsa-miR-542-5p, hsa-miR-502-3p, hsa-miR-376a, hsa-miR-500a, hsa-miR-424) were insensitive to perturbations according to jackknife cross-validation on their HR for OS. The respective principal component analysis (PCA) defined a meta-miRNA signature including the same 6 miRNA's, resulting in a HR of 0.66 (95%-CI 0.53-0.82). CONCLUSION Cell-free circulating miRNA-profiling successfully identified a highly prognostic 6-gene signature in patients with advanced non-squamous NSCLC. Circulating miRNA profiling should further be validated in external cohorts for the selection and monitoring of systemic treatment in patients with advanced NSCLC.

Coherence and Elicitability

The risk of a financial position is usually summarized by a risk measure. As this risk measure has to be estimated from historical data, it is important to be able to verify and compare competing estimation procedures. In statistical decision theory, risk measures for which such verification and comparison is possible, are called elicitable. It is known that quantile-based risk measures such as value at risk are elicitable. In this paper, the existing result of the nonelicitability of expected shortfall is extended to all law-invariant spectral risk measures unless they reduce to minus the expected value. Hence, it is unclear how to perform forecast verification or comparison. However, the class of elicitable law-invariant coherent risk measures does not reduce to minus the expected value. We show that it consists of certain expectiles.

A simulation study on the statistical monitoring of condemnation rates from slaughterhouses for syndromic surveillance: an evaluation based on Swiss data

Syndromic surveillance (SyS) systems currently exploit various sources of health-related data, most of which are collected for purposes other than surveillance (e.g. economic). Several European SyS systems use data collected during meat inspection for syndromic surveillance of animal health, as some diseases may be more easily detected post-mortem than at their point of origin or during the ante-mortem inspection upon arrival at the slaughterhouse. In this paper we use simulation to evaluate the performance of a quasi-Poisson regression (also known as an improved Farrington) algorithm for the detection of disease outbreaks during post-mortem inspection of slaughtered animals. When parameterizing the algorithm based on the retrospective analyses of 6 years of historic data, the probability of detection was satisfactory for large (range 83-445 cases) outbreaks but poor for small (range 20-177 cases) outbreaks. Varying the amount of historical data used to fit the algorithm can help increasing the probability of detection for small outbreaks. However, while the use of a 0·975 quantile generated a low false-positive rate, in most cases, more than 50% of outbreak cases had already occurred at the time of detection. High variance observed in the whole carcass condemnations time-series, and lack of flexibility in terms of the temporal distribution of simulated outbreaks resulting from low reporting frequency (monthly), constitute major challenges for early detection of outbreaks in the livestock population based on meat inspection data. Reporting frequency should be increased in the future to improve timeliness of the SyS system while increased sensitivity may be achieved by integrating meat inspection data into a multivariate system simultaneously evaluating multiple sources of data on livestock health.

MOREMATA: Stata module (Mata) to provide various functions

This package includes various Mata functions. kern(): various kernel functions; kint(): kernel integral functions; kdel0(): canonical bandwidth of kernel; quantile(): quantile function; median(): median; iqrange(): inter-quartile range; ecdf(): cumulative distribution function; relrank(): grade transformation; ranks(): ranks/cumulative frequencies; freq(): compute frequency counts; histogram(): produce histogram data; mgof(): multinomial goodness-of-fit tests; collapse(): summary statistics by subgroups; _collapse(): summary statistics by subgroups; gini(): Gini coefficient; sample(): draw random sample; srswr(): SRS with replacement; srswor(): SRS without replacement; upswr(): UPS with replacement; upswor(): UPS without replacement; bs(): bootstrap estimation; bs2(): bootstrap estimation; bs_report(): report bootstrap results; jk(): jackknife estimation; jk_report(): report jackknife results; subset(): obtain subsets, one at a time; composition(): obtain compositions, one by one; ncompositions(): determine number of compositions; partition(): obtain partitions, one at a time; npartitionss(): determine number of partitions; rsubset(): draw random subset; rcomposition(): draw random composition; colvar(): variance, by column; meancolvar(): mean and variance, by column; variance0(): population variance; meanvariance0(): mean and population variance; mse(): mean squared error; colmse(): mean squared error, by column; sse(): sum of squared errors; colsse(): sum of squared errors, by column; benford(): Benford distribution; cauchy(): cumulative Cauchy-Lorentz dist.; cauchyden(): Cauchy-Lorentz density; cauchytail(): reverse cumulative Cauchy-Lorentz; invcauchy(): inverse cumulative Cauchy-Lorentz; rbinomial(): generate binomial random numbers; cebinomial(): cond. expect. of binomial r.v.; root(): Brent's univariate zero finder; nrroot(): Newton-Raphson zero finder; finvert(): univariate function inverter; integrate_sr(): univariate function integration (Simpson's rule); integrate_38(): univariate function integration (Simpson's 3/8 rule); ipolate(): linear interpolation; polint(): polynomial inter-/extrapolation; plot(): Draw twoway plot; _plot(): Draw twoway plot; panels(): identify nested panel structure; _panels(): identify panel sizes; npanels(): identify number of panels; nunique(): count number of distinct values; nuniqrows(): count number of unique rows; isconstant(): whether matrix is constant; nobs(): number of observations; colrunsum(): running sum of each column; linbin(): linear binning; fastlinbin(): fast linear binning; exactbin(): exact binning; makegrid(): equally spaced grid points; cut(): categorize data vector; posof(): find element in vector; which(): positions of nonzero elements; locate(): search an ordered vector; hunt(): consecutive search; cond(): matrix conditional operator; expand(): duplicate single rows/columns; _expand(): duplicate rows/columns in place; repeat(): duplicate contents as a whole; _repeat(): duplicate contents in place; unorder2(): stable version of unorder(); jumble2(): stable version of jumble(); _jumble2(): stable version of _jumble(); pieces(): break string into pieces; npieces(): count number of pieces; _npieces(): count number of pieces; invtokens(): reverse of tokens(); realofstr(): convert string into real; strexpand(): expand string argument; matlist(): display a (real) matrix; insheet(): read spreadsheet file; infile(): read free-format file; outsheet(): write spreadsheet file; callf(): pass optional args to function; callf_setup(): setup for mm_callf().