Inferential, non-parametric statistics to assess quality of probabilistic forecast systems

Many statistical forecast systems are available to interested users. In order to be useful for decision-making, these systems must be based on evidence of underlying mechanisms. Once causal connections between the mechanism and their statistical manifestation have been firmly established, the forecasts must also provide some quantitative evidence of `quality’. However, the quality of statistical climate forecast systems (forecast quality) is an ill-defined and frequently misunderstood property. Often, providers and users of such forecast systems are unclear about what ‘quality’ entails and how to measure it, leading to confusion and misinformation. Here we present a generic framework to quantify aspects of forecast quality using an inferential approach to calculate nominal significance levels (p-values) that can be obtained either by directly applying non-parametric statistical tests such as Kruskal-Wallis (KW) or Kolmogorov-Smirnov (KS) or by using Monte-Carlo methods (in the case of forecast skill scores). Once converted to p-values, these forecast quality measures provide a means to objectively evaluate and compare temporal and spatial patterns of forecast quality across datasets and forecast systems. Our analysis demonstrates the importance of providing p-values rather than adopting some arbitrarily chosen significance levels such as p < 0.05 or p < 0.01, which is still common practice. This is illustrated by applying non-parametric tests (such as KW and KS) and skill scoring methods (LEPS and RPSS) to the 5-phase Southern Oscillation Index classification system using historical rainfall data from Australia, The Republic of South Africa and India. The selection of quality measures is solely based on their common use and does not constitute endorsement. We found that non-parametric statistical tests can be adequate proxies for skill measures such as LEPS or RPSS. The framework can be implemented anywhere, regardless of dataset, forecast system or quality measure. Eventually such inferential evidence should be complimented by descriptive statistical methods in order to fully assist in operational risk management.

Strong evidence for a novel schizophrenia risk locus on chromosome 1p31.1 in homogeneous pedigrees from Tamil Nadu, India

OBJECTIVE: The study of ethnically homogeneous populations may help to identify schizophrenia risk loci. The authors conducted a genomewide linkage scan for schizophrenia in an Indian population. METHOD: Participants were 441 individuals (262 affected probands and siblings) who were recruited primarily from one ethnically homogeneous group, the Tamil Brahmin caste, although individuals from other geographically proximal castes also participated. Genotyping of 124 affected sibling pair pedigrees was performed with 402 short tandem repeat polymorphisms. Linkage analyses were conducted using nonparametric exponential LOD (logarithm of the odds ratio for linkage) scores and parametric heterogeneity LOD scores. Parametric heterogeneity scores were calculated using simple dominant and recessive models, correcting for multiple statistics. The data were examined for evidence of consanguinity. Genomewide significance levels were determined using 10,000 gene dropping simulations. RESULTS: These findings revealed genomewide significant linkage to chromosome 1p31.1, through the use of both exponential and heterogeneity LOD scores, incorporating correction for multiple statistics and mild consanguinity. The estimated sibling recurrence risk associated with this putative locus was 1.95. Analysis for heterogeneity LOD scores also detected suggestive linkage to chromosomes 13q22.1 and 16q12.2. Using 117 tag single nucleotide polymorphisms (SNPs), family-based association analyses of phosphodiesterase 4B (PDE4B), the closest schizophrenia candidate gene, detected no convincing evidence of association, suggesting that the chromosome 1 peak represents a novel risk locus. CONCLUSIONS: This is the first study-to the authors' knowledge-to report significant linkage of schizophrenia to chromosome 1p31.1. Further investigation of this chromosome region in diverse populations is warranted to identify underlying sequence variants.

A genome-wide linkage scan provides evidence for both new and previously reported loci influencing common migraine

Latent class analysis was performed on migraine symptom data collected in a Dutch population sample (N = 12,210, 59% female) in order to obtain empirical groupings of individuals suffering from symptoms of migraine headache. Based on these heritable groupings (h(2) = 0.49, 95% CI: 0.41-0.57) individuals were classified as affected (migrainous headache) or unaffected. Genome-wide linkage analysis was performed using genotype data from 105 families with at least 2 affected siblings. In addition to this primary phenotype, linkage analyses were performed for the individual migraine symptoms. Significance levels, corrected for the analysis of multiple traits, were determined empirically via a novel simulation approach. Suggestive linkage for migrainous headache was found on chromosomes 1 (LOD = 1.63; pointwise P = 0.0031), 13 (LOD = 1.63; P = 0.0031), and 20 (LOD = 1.85; P = 0.0018). Interestingly, the chromosome 1 peak was located close to the ATP1A2 gene, associated with familial hemiplegic migraine type 2 (FHM2). Individual symptom analysis produced a LOD score of 1.97 (P = 0.0013) on chromosome 5 (photo/phonophobia), a LOD score of 2.13 (P = 0.0009) on chromosome 10 (moderate/severe pain intensity) and a near significant LOD score of 3.31 (P = 0.00005) on chromosome 13 (pulsating headache). These peaks were all located near regions previously reported in migraine linkage studies. Our results provide important replication and support for the presence of migraine susceptibility genes within these regions, and further support the utility of an LCA-based phenotyping approach and analysis of individual symptoms in migraine genetic research. Additionally, our novel "2-step" analysis and simulation approach provides a powerful means to investigate linkage to individual trait components.

A reanalysis of 409 European-Ancestry and African American schizophrenia pedigrees reveals significant linkage to 8p23.3 with evidence of locus heterogeneity

The detection and replication of schizophrenia risk loci can require substantial sample sizes, which has prompted various collaborative efforts for combining multiple samples. However, pooled samples may comprise sub-samples with substantial population genetic differences, including allele frequency differences. We investigated the impact of population differences via linkage reanalysis of Molecular Genetics of Schizophrenia 1 (MGS1) affected sibling-pair data, comprising two samples of distinct ancestral origin: European (EA: 263 pedigrees) and African-American (AA: 146 pedigrees). To exploit the linkage information contained within these distinct continental samples, we performed separate analyses of the individual samples, allowing for within-sample locus heterogeneity, and the pooled sample, allowing for both within-sample and between-sample heterogeneity. Significance levels, corrected for the multiple tests, were determined empirically. For all suggestive peaks, stronger linkage evidence was obtained in either the EA or AA sample than the combined sample, regardless of how heterogeneity was modeled for the latter. Notably, we report genomewide significant linkage of schizophrenia to 8p23.3 and evidence for a second, independent susceptibility locus, reaching suggestive linkage, 29 cM away on 8p21.3. We also detected suggestive linkage on chromosomes 5p13.3 and 7q36.2. Many regions showed pronounced differences in the extent of linkage between the EA and AA samples. This reanalysis highlights the potential impact of population differences upon linkage evidence in pooled data and demonstrates a useful approach for the analysis of samples drawn from distinct continental groups.

Análise comparativa dos modelos de cálculo dos retornos anormais utilizando o evento recompra de ações na Bovespa

Este trabalho tem por objetivo verificar se há diferença quanto ao nível de significância estatística no cálculo do retorno anormal realizado através de quatro modelos estatísticos utilizados em estudos de eventos, tendo como objeto de estudo empresas no mercado de ações no Brasil durante o período de março de 2003 até julho de 2010 na Bovespa. Considerando a importância do cálculo do retorno anormal nos estudos de eventos, bem como possíveis variações de resultados decorrentes da escolha do modelo de cálculo, este estudo utiliza um tema bastante conhecido, qual seja, o anúncio de recompra de ações feito pela própria companhia emissora dos títulos. A metodologia utilizada nesta pesquisa foi quantitativa, utilizando o estudo de corte transversal. Os resultados apontam que há diferença entre os níveis de significância encontrados. Ao analisar o gráfico dos modelos calculados no período da janela do evento, verificou-se que as empresas que recompraram ações a fizeram quando os papéis estavam com retorno anormal acumulado negativo e, após a recompra, os papéis tiveram retornos anormais acumulados positivos. Recalculou-se os dois modelos que utilizam o Ibovespa em sua fórmula de cálculo, através de um Ibovespa sem ponderação e conclui-se que os resultados apontam na direção de se evitar o uso de índices ponderados de mercado, preferindo a utilização de carteiras compostas apenas com uma ação para cada empresa componente da carteira de controle. Após este recálculo, verificou-se que o modelo que era menos próximo dos demais graficamente era o modelo de retorno ajustado ao mercado ponderado. Por fim, as evidências empíricas indicam que o mercado de capitais brasileiro ajusta tempestivamente os papéis das empresas que realizaram recompra de ações, em linha com o que prescreve a hipótese do mercado eficiente na sua forma semiforte.

Twisting flux tubes as a cause of micro-flaring activity

High-cadence optical observations of an H-alpha blue-wing bright point near solar AR NOAA 10794 are presented. The data were obtained with the Dunn Solar Telescope at the National Solar Observatory/Sacramento Peak using a newly developed camera system, the RAPID DUAL IMAGER. Wavelet analysis is undertaken to search for intensity-related oscillatory signatures, and periodicities ranging from 15 to 370 s are found with significance levels exceeding 95%. During two separate microflaring events, oscillation sites surrounding the bright point are observed to twist. We relate the twisting of the oscillation sites to the twisting of physical flux tubes, thus giving rise to reconnection phenomena. We derive an average twist velocity of 8.1 km s(-1) and detect a peak in the emitted flux between twist angles of 180 degrees and 230 degrees.

High-frequency oscillations in a solar active region observed with the RAPID DUAL IMAGER

High-cadence, synchronized, multiwavelength optical observations of a solar active region (NOAA 10794) are presented. The data were obtained with the Dunn Solar Telescope at the National Solar Observatory/Sacramento Peak using a newly developed camera system: the rapid dual imager. Wavelet analysis is undertaken to search for intensity related oscillatory signatures, and periodicities ranging from 20 to 370 s are found with significance levels exceeding 95%. Observations in the H-α blue wing show more penumbral oscillatory phenomena when compared to simultaneous G-band observations. The H-α oscillations are interpreted as the signatures of plasma motions with a mean velocity of 20 km s-1. The strong oscillatory power over H-α blue-wing and G-band penumbral bright grains is an indication of the Evershed flow with frequencies higher than previously reported.

Alfvén Waves in the Lower Solar Atmosphere

The flow of energy through the solar atmosphere and the heating of the Sun's outer regions are still not understood. Here, we report the detection of oscillatory phenomena associated with a large bright-point group that is 430,000 square kilometers in area and located near the solar disk center. Wavelet analysis reveals full-width half-maximum oscillations with periodicities ranging from 126 to 700 seconds originating above the bright point and significance levels exceeding 99%. These oscillations, 2.6 kilometers per second in amplitude, are coupled with chromospheric line-of-sight Doppler velocities with an average blue shift of 23 kilometers per second. A lack of cospatial intensity oscillations and transversal displacements rules out the presence of magneto-acoustic wave modes. The oscillations are a signature of Alfvén waves produced by a torsional twist of ±22 degrees. A phase shift of 180 degrees across the diameter of the bright point suggests that these torsional Alfvén oscillations are induced globally throughout the entire brightening. The energy flux associated with this wave mode is sufficient to heat the solar corona.

Preferred structures in large-scale circulation and the effect of doubling greenhouse gas concentration in HadCM3

Preferred structures in the surface pressure variability are investigated in and compared between two 100-year simulations of the Hadley Centre climate model HadCM3. In the first (control) simulation, the model is forced with pre-industrial carbon dioxide concentration (1×CO2) and in the second simulation the model is forced with doubled CO2 concentration (2×CO2). Daily winter (December-January-February) surface pressures over the Northern Hemisphere are analysed. The identification of preferred patterns is addressed using multivariate mixture models. For the control simulation, two significant flow regimes are obtained at 5% and 2.5% significance levels within the state space spanned by the leading two principal components. They show a high pressure centre over the North Pacific/Aleutian Islands associated with a low pressure centre over the North Atlantic, and its reverse. For the 2×CO2 simulation, no such behaviour is obtained. At higher-dimensional state space, flow patterns are obtained from both simulations. They are found to be significant at the 1% level for the control simulation and at the 2.5% level for the 2×CO2 simulation. Hence under CO2 doubling, regime behaviour in the large-scale wave dynamics weakens. Doubling greenhouse gas concentration affects both the frequency of occurrence of regimes and also the pattern structures. The less frequent regime becomes amplified and the more frequent regime weakens. The largest change is observed over the Pacific where a significant deepening of the Aleutian low is obtained under CO2 doubling.

Forecasting VaR using analytic higher moments for GARCH processes

It is widely accepted that some of the most accurate Value-at-Risk (VaR) estimates are based on an appropriately specified GARCH process. But when the forecast horizon is greater than the frequency of the GARCH model, such predictions have typically required time-consuming simulations of the aggregated returns distributions. This paper shows that fast, quasi-analytic GARCH VaR calculations can be based on new formulae for the first four moments of aggregated GARCH returns. Our extensive empirical study compares the Cornish–Fisher expansion with the Johnson SU distribution for fitting distributions to analytic moments of normal and Student t, symmetric and asymmetric (GJR) GARCH processes to returns data on different financial assets, for the purpose of deriving accurate GARCH VaR forecasts over multiple horizons and significance levels.

Twenty-three cycles of changing open solar magnetic flux

This paper presents a comparison of various estimates of the open solar flux, deduced from measurements of the interplanetary magnetic field, from the aa geomagnetic index and from photospheric magnetic field observations. The first two of these estimates are made using the Ulysses discovery that the radial heliospheric field is approximately independent of heliographic latitude, the third makes use of the potential-field source surface method to map the total flux through the photosphere to the open flux at the top of the corona. The uncertainties associated with using the Ulysses result are 5%, but the effects of the assumptions of the potential field source surface method are harder to evaluate. Nevertheless, the three methods give similar results for the last three solar cycles when the data sets overlap. In 11-year running means, all three methods reveal that 1987 marked a significant peak in the long-term variation of the open solar flux. This peak is close to the solar minimum between sunspot cycles 21 and 22, and consequently the mean open flux (averaged from minimum to minimum) is similar for these two cycles. However, this similarity between cycles 21 and 22 in no way implies that the open flux is constant. The long-term variation shows that these cycles are fundamentally different in that the average open flux was rising during cycle 21 (from consistently lower values in cycle 20 and toward the peak in 1987) but was falling during cycle 22 (toward consistently lower values in cycle 23). The estimates from the geomagnetic aa index are unique as they extend from 1842 onwards (using the Helsinki extension). This variation gives strong anticorrelations, with very high statistical significance levels, with cosmic ray fluxes and with the abundances of the cosmogenic isotopes that they produce. Thus observations of photospheric magnetic fields, of cosmic ray fluxes, and of cosmogenic isotope abundances all support the long-term drifts in open solar flux reported by Lockwood et al. [1999a, 1999b].

An evaluation of the correlation between open solar flux and total solar irradiance

The correlation between the coronal source flux F_{S} and the total solar irradiance I_{TS} is re-evaluated in the light of an additional 5 years' data from the rising phase of solar cycle 23 and also by using cosmic ray fluxes detected at Earth. Tests on monthly averages show that the correlation with F_{S} deduced from the interplanetary magnetic field (correlation coefficient, r = 0.62) is highly significant (99.999%), but that there is insufficient data for the higher correlation with annual means (r = 0.80) to be considered significant. Anti-correlations between I_{TS} and cosmic ray fluxes are found in monthly data for all stations and geomagnetic rigidity cut-offs (r ranging from −0.63 to −0.74) and these have significance levels between 85% and 98%. In all cases, the t is poorest for the earliest data (i.e., prior to 1982). Excluding these data improves the anticorrelation with cosmic rays to r = −0:93 for one-year running means. Both the interplanetary magnetic field data and the cosmic ray fluxes indicate that the total solar irradiance lags behind the open solar flux with a delay that is estimated to have an optimum value of 2.8 months (and is within the uncertainty range 0.8-8.0 months at the 90% level).

Robust inference in an heteroscedastic measurement error model

In this paper we deal with robust inference in heteroscedastic measurement error models Rather than the normal distribution we postulate a Student t distribution for the observed variables Maximum likelihood estimates are computed numerically Consistent estimation of the asymptotic covariance matrices of the maximum likelihood and generalized least squares estimators is also discussed Three test statistics are proposed for testing hypotheses of interest with the asymptotic chi-square distribution which guarantees correct asymptotic significance levels Results of simulations and an application to a real data set are also reported (C) 2009 The Korean Statistical Society Published by Elsevier B V All rights reserved

Bond Strength of Dental Adhesive Systems Irradiated with Ionizing Radiation

Purpose: The aim of the present paper was to determine the effect of different types of ionizing radiation on the bond strength of three different dentin adhesive systems. Materials and Methods: One hundred twenty specimens of 60 human teeth (protocol number: 032/2007) sectioned mesiodistally were divided into 3 groups according to the adhesives systems used: SB (Adper Single Bond Plus), CB (Clearfil SE Bond) and AP (Adper Prompt Self-Etch). The adhesives were applied on dentin and photo-activated using LED (Lec 1000, MMoptics, 1000 mW/cm(2)). Customized elastomer molds (0.5 mm thickness) with three orifices of 1.2 mm diameter were placed onto the bonding areas and filled with composite resin (Filtek Z-250), which was photo-activated for 20 s. Each group was subdivided into 4 Subgroups for application of the different types of ionizing radiation: ultraviolet radiation (UV), diagnostic x-ray radiation (DX), therapeutic x-ray radiation (TX) and without irradiation (control group, CG). Microshear tests were carried out (Instron, model 4411), and afterwards the modes of failure were evaluated by optical and scanning electron microscope and classified using 5 scores: adhesive failure, mixed failures with 3 significance levels, and cohesive failure. The results of the shear bond strength test were submitted to ANOVA with Tukey`s test and Dunnett`s test, and the data from the failure pattern evaluation were analyzed with the Mann Whitney test (p = 0.05). Results: No change in bond strength of CB and AP was observed after application of the different radiation types, only SB showed increase in bond strength after UV (p = 0.0267) irradiation. The UV also changed the failure patterns of SB (p = 0.0001). Conclusion: The radio-induced changes did not cause degradation of the restorations, which means that they can be exposed to these types of ionizing radiation without weakening the bond strength.

Comparação de Diferentes Modelos para Avaliação Genética de Características de Desempenho Pós-desmama em Suínos

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)