Choice Of Threshold Research Articles

Cross-Validatory Extreme Value Threshold Selection and Uncertainty with Application to Ocean Storm Severity

SummaryDesign conditions for marine structures are typically informed by threshold-based extreme value analyses of oceanographic variables, in which excesses of a high threshold are modelled by a generalized Pareto distribution. Too low a threshold leads to bias from model misspecification, and raising the threshold increases the variance of estimators: a bias–variance trade-off. Many existing threshold selection methods do not address this trade-off directly but rather aim to select the lowest threshold above which the generalized Pareto model is judged to hold approximately. In the paper Bayesian cross-validation is used to address the trade-off by comparing thresholds based on predictive ability at extreme levels. Extremal inferences can be sensitive to the choice of a single threshold. We use Bayesian model averaging to combine inferences from many thresholds, thereby reducing sensitivity to the choice of a single threshold. The methodology is applied to significant wave height data sets from the northern North Sea and the Gulf of Mexico.

Large scale spectrum survey in rural and urban environments within the 50 MHz–6 GHz bands

In-depth spectrum measurement was conducted in rural and urban locations, covering 50MHz–6GHz bands, during the weekdays and weekends. A modified duty cycle metric is presented by introducing a space variable into the existing metrics available today. An adaptive energy detection threshold technique was employed, the results indicate the average spectral occupancy of 0.18%, and 5.08% in rural and urban locations respectively during weekdays and 1.45% on weekends for urban locations. Furthermore, short and long term temporal variations of the duty cycle for each of the bands were studied, and it was found that GSM 900 shows significant temporal variation when compared with GSM 1800. It was also found that the choice of the detection threshold would significantly affect the duty cycle as GSM 900 and 1800 give exponential decay with increase in detection threshold while TV band shows very sharp exponential decay which becomes invariant after −85dBm.

Cost-effectiveness of implant-supported mandibular removable partial dentures.

The aim of this study was to conduct a cost-effectiveness analysis comparing conventional removable partial dentures (RPDs) and implant-supported RPDs (ISRPDs) treatment in patients with an edentulous maxilla and a bilateral free-ending situation in the mandible. Thirty subjects were included. A new RPD was made and implant support was provided 3months later. Treatment costs (opportunity costs and costs based on tariffs) were calculated. Treatment effect was expressed by means of the Dutch Oral Health Impact Profile questionnaire (OHIP-NL49), a chewing ability test (Mixing Ability Index, MAI) and a short-form health survey measuring perceived general health (SF-36), which was subsequently converted into quality-adjusted-life-years (QALYs). The incremental cost-effectiveness ratio (ICER) was the primary outcome measure of cost-effectiveness, comparing both treatment strategies. The mean total opportunity costs were €981 (95% CI €971-€991) for the RPD treatment and €2.480 (95% CI €2.461-€2.500) for the ISRPD treatment. The total costs derived from the national tariff structure were €850 for the RPD treatment and €2.610 for the ISRPD treatment. The ICER for OHIP-NL49 and MAI using the opportunity costs was €80 and €786, respectively. When using the tariff structure, corresponding ICERs were €94 and €921. The effect of supporting an RPD with implants when expressed in QALYs was negligible; hence an ICER was not determined. It is concluded that depending on the choice of outcome measure and monetary threshold, supporting an RPD with implants is cost-effective when payers are willing to pay more than €80 per OHIP point gained. Per MAI point gained, an additional €786 has to be invested.

Risk factors for residual obstructive sleep apnea after adenotonsillectomy in children.

To determine the prevalence of residual obstructive sleep apnea (OSA) in children who had adenotonsillectomy (AT) and to identify the risk factors for residual OSA after AT. Retrospective chart review. Children with OSA who had AT at a tertiary care children's hospital were reviewed. Data pertaining to demographics, past medical history, body mass index, tonsil and adenoid size, and polysomnography were obtained. Residual OSA was defined as apnea hypopnea index (AHI) greater than 2. The rate of residual OSA and risk factors for residual OSA were assessed. One hundred sixty-nine children with OSA underwent polysomnography before and after AT. The prevalence of residual OSA was 38%. The prevalence of residual OSA in obese patients (49%) was higher than that of nonobese patients (27%) (P = .02). Patients with neurological/developmental/craniofacial abnormalities had higher prevalence of residual OSA (44%) than patients without comorbidities (33%) (P < .05). The prevalence of residual OSA in patients with severe OSA (42%) was higher than patients with moderate (29%) or mild OSA (0%) (P = .03). Teenage patients (67%) had a higher prevalence of residual OSA than toddlers (27%), preschooler (33%), and middle childhood groups (29%) (P = .03). The majority of children had improvement in OSA after AT. The choice of AHI threshold used to define residual OSA influenced the prevalence of residual OSA. Teenagers and children with obesity, comorbidities including neurological/developmental/craniofacial abnormalities alone or in combination with asthma, or severe OSA have a high risk of residual OSA. 4 Laryngoscope, 126:2624-2629, 2016.

The challenge of predicting currency crises: how do definition and probability threshold choice make a difference?

In order to analyze the definition effect on prediction powers of Early Warning Systems (EWS), we construct an EWS model with 20 different crisis definitions. Furthermore, to analyze the probability threshold effect, we identify 11 different threshold levels and forecast the different versions of our model for each of those threshold levels. Our results show that crisis definitions and threshold choices significantly affect the prediction powers of the EWS models. To put it more explicitly, Exchange Market Pressure (EMP) Index based definition is shown to be a better predictor compared to depreciation based definition. Furthermore, EMP Index is found to give better results with higher standard deviation multiplier. Last but not least, it is empirically proven that 50 % threshold is the optimal level for EWS analyses as until that level the prediction powers of the models significantly increase but keep constant above it.

Maremar, prevalence of chronic kidney disease, how to avoid over-diagnosis and under-diagnosis

Chronic kidney disease is considered as a major public health problem. Recent studies mention a prevalence rate between 8%–12%. Several editorials, comments, short reviews described the weaknesses (lack of confirmation of proteinuria, and of chronicity of decreased estimated glomerular filtration rate) of a substantial number of studies and the irrational of using a single arbitrary set point, i.e. diagnosis of chronic kidney disease whenever the estimated glomerular filtration rate is less than 60mL/min/1.73m2. Maremar (Maladies rénales chroniques au Maroc) is a prevalence study of chronic kidney disease, hypertension, diabetes and obesity in a randomized, representative, high response rate (85%), sample of the adult population of Morocco, strictly applying the Kidney Disease: Improving Global Outcomes (KDIGO) criteria. Compared to the vast majority of the available studies, Maremar has a low prevalence of chronic kidney disease (2.9% adjusted to the actual adult population of Morocco). The population pyramid, and particularly the confirmation of proteinuria and “chronicity” of the decreased estimated glomerular filtration rate are the main reasons for this low prevalence of chronic kidney disease. The choice of arbitrary single threshold of estimated glomerular filtration rate for classifying stage 3–5 chronic kidney disease inevitably leads to “over-diagnosis” (false positives) of the disease in the elderly, particularly those without proteinuria, hematuria or hypertension, and to “under-diagnosed” (false negatives) in younger individuals with an estimated glomerular filtration rate above 60mL/min/1.73m2 and below the 3rd percentile of their age/gender category. There is an urgent need for quality studies using in a correct way the recent KDIGO guidelines when investigating the prevalence of chronic kidney disease, in order to avoid a 50 to 100% overestimation of a disease state with potential dramatic consequences. The combination of the general population screening encompassing four different major health problems in the same screening procedure, using the correct methodologies and procedures, combined with a prevention/follow-up program results in a clinically/scientifically relevant program.

Maximum Wind Gusts Associated with Human-Reported Nonconvective Wind Events and a Comparison to Current Warning Issuance Criteria

Abstract Nonconvective high winds are a deceptively hazardous meteorological phenomenon. Though the National Weather Service (NWS) possesses an array of products designed to alert the public to nonconvective wind potential, documentation justifying the choice of issuance thresholds is scarce. Measured wind speeds from the Global Historical Climatology Network (GHCN)-Daily dataset associated with human-reported nonconvective wind events from Storm Data are examined in order to assess the suitability of the current gust criteria for the NWS wind advisory and high wind warning. Nearly 92% (45%) of the nonconvective wind events considered from Storm Data were accompanied by peak gusts beneath the high wind warning (wind advisory) threshold of 58 mi h−1 (25.9 m s−1) [46 mi h−1 (20.6 m s−1)], and greater than 74% (28%) of all fatal and injury-causing events were associated with peak gusts below these same product gust criteria. NWS wind products were disproportionately issued in areas of complex terrain where wind climatologies include a greater frequency of high wind warning threshold-level gusts, irrespective of observed impacts. For many areas of the eastern United States, a 58 mi h−1 (25.9 m s−1) gust of convective, tropical, or nonconvective origin falls within the top 0.5% of all observed daily maximum wind gusts, nearly eliminating the possibility of a nonconvective gust meeting the issuance criterion.

Assessing minimum environmental flows in nonpermanent rivers: The choice of thresholds

The criteria used in the computation of the minimum environmental flow regime and flow cessation periods in nonpermanent rivers are often left to open criteria. This study proposes a stochastic approach for evaluating the choice of local thresholds in the characterization of minimum environmental flows through both the Monte Carlo technique and local hydrological relationships. This approach is applied to four regimes obtained by hydrologic and hydraulic habitat modeling in a Mediterranean watershed. The operationality, defined as the probability of the calculated environmental regime being satisfied by the natural regime over 25 years, was assessed for eight different scenarios. Two monthly minimum environmental flow regimes were then generated, with 90 and 95% operationality levels. This analysis allows the generation of minimum flow regime prescriptions from a strictly hydrologic point of view. The methodology proposed constitutes a useful tool for the implementation of uncertainty analysis of environmental flows in water resource management.

Energy landscapes and persistent minima.

We consider a coarse-graining of high-dimensional potential energy landscapes based upon persistences, which correspond to lowest barrier heights to lower-energy minima. Persistences can be calculated efficiently for local minima in kinetic transition networks that are based on stationary points of the prevailing energy landscape. The networks studied here represent peptides, proteins, nucleic acids, an atomic cluster, and a glassy system. Minima with high persistence values are likely to represent some form of alternative structural morphology, which, if appreciably populated at the prevailing temperature, could compete with the global minimum (defined as infinitely persistent). Threshold values on persistences (and in some cases equilibrium occupation probabilities) have therefore been used in this work to select subsets of minima, which were then analysed to see how well they can represent features of the full network. Simplified disconnectivity graphs showing only the selected minima can convey the funnelling (including any multiple-funnel) characteristics of the corresponding full graphs. The effect of the choice of persistence threshold on the reduced disconnectivity graphs was considered for a system with a hierarchical, glassy landscape. Sets of persistent minima were also found to be useful in comparing networks for the same system sampled under different conditions, using minimum oriented spanning forests.

Adaptive one-bit quantization for compressed sensing

There have been a number of studies on sparse signal recovery from one-bit quantized measurements. Nevertheless, less attention has been paid to the choice of the quantization thresholds and its impact on the signal recovery performance. In this paper, we examine the problem of quantization in a general framework of one-bit compressed sensing with non-zero quantization thresholds. Our analysis shows that when the number of one-bit measurements is sufficiently large, with a high probability the sparse signal can be recovered with an error decaying linearly with the ℓ2-norm of the difference between the quantization thresholds and the original unquantized measurements. Specifically, by setting the thresholds sufficiently close to the original unquantized measurements, sparse signals can be recovered with an arbitrarily small error. By borrowing an idea from the Delta modulation, we propose an adaptive quantization scheme where the quantization thresholds are iteratively adjusted based on previous encoded bits such that they eventually oscillate around the original unquantized measurements with decreasing granular noise. Numerical results are provided to collaborate our theoretical results and to illustrate the effectiveness of the proposed scheme.

Digital filtering of noisy data concerned with the effect of the threshold level and choice of a wavelet

The problem of digital wavelet filtering of noisy data is considered. The results are comparatively analyzed when the threshold function is specified in different ways for wavelet transformation coefficients. The influence of the choice of the wavelet basis on the efficiency of the noise removal is studied. It is shown that the minimum filtering error is in most cases provided when the soft variant of the threshold function introduction is introduced.

On the integration of input and output control: Workload Control order release

Workload Control is a production planning and control concept developed for high-variety job shops. It integrates two control mechanisms: (i) input control, to regulate the inflow of work to the system; and (ii) output control, which uses capacity adjustments to regulate the outflow of work from the system. Much Workload Control research has focused on input control, while output control has been largely neglected. Only recently has research emerged that uses Workload Control theory to guide capacity adjustments. Yet this literature focuses on capacity adjustments (output control) only – it fails to integrate it with Workload Control’s input control element. In response, this study explores the performance impact of Workload Control when input control (controlled order release) and output control (capacity adjustments) are combined. Job shop simulation results demonstrate that input and output control can and should play complementary roles. Both elements significantly enhance performance in isolation, and performance effects appear to complement each other. Further, results indicate that the choice of the workload threshold that triggers capacity adjustments has a stronger impact on performance than the actual size of the adjustment. The measure of workload used to guide the load-based order release decision is also used to determine the workload threshold that triggers the capacity adjustment. This facilitates implementation in practice. Finally, although our study is on Workload Control, the findings have important implications for other production planning and control concepts.

Catastrophic medical expenditure risk

We propose a measure of household exposure to particularly onerous medical expenses. The measure can be decomposed into the probability that medical expenditure exceeds a threshold, the loss due to predictably low consumption of other goods if it does and the further loss arising from the volatility of medical expenses above the threshold. Depending on the choice of threshold, the measure is consistent with a model of reference-dependent utility with loss aversion. Unlike the risk premium, the measure is only sensitive to particularly high expenses, and can identify households that expect to incur such expenses and would benefit from subsidised, but not actuarially fair, insurance. An empirical illustration using data from seven Asian countries demonstrates the importance of taking account of informal insurance and reveals clear differences in catastrophic medical expenditure risk across and within countries. In general, risk is higher among poorer, rural and chronically ill populations.

Exponentiality Test Procedures for Large Samples of Rainfall Event Characteristics

AbstractThe main purpose of this paper is to examine and recommend procedures that can be used to statistically test the exponentiality of large amounts of sample data for rainfall event volume, duration, and interevent time. Based on literature review and initial analysis, the Poisson and chi-square goodness-of-fit tests are selected first. Some misconceptions about parameter estimators and degrees of freedom associated with the use of the chi-square goodness-of-fit tests are then clarified. Using rainfall data from seven stations in the north-central region of the United States, the choice of the event volume threshold and the minimum interevent time for separating continuous rainfall data into individual events are examined in detail. Findings from this study suggest that the Poisson test can be used for testing the exponentiality of interevent times and for examining the statistical independence of consecutive rainfall events. The use of the minimum chi-square estimator combined with the chi-square go...

Advanced Blanking Nonlinearity for Mitigating Impulsive Interference in OFDM Systems

In this paper, we introduce advancements of the conventional blanking nonlinearity (BN) for orthogonal frequency-division multiplexing (OFDM)-based systems, which is referred to in the following as advanced BN. Blanking is a common measure for mitigating impulsive interference that often occurs in wireless communication systems. Although the BN removes impulsive interference reliably, it possesses various drawbacks for OFDM-based systems. In particular, the choice of the blanking threshold (BT), to decide whether a received sample is blanked, is a critical issue. We present an algorithm for determining the optimal BT to maximize the signal-to-noise-and-interference ratio (SINR) after blanking. Another drawback is that the entire received signal is discarded during a blanking interval, despite the fact that only a fraction of the spectrum of the OFDM signal might be affected by interference. We show how blanking can be limited to subcarriers that are actually affected by interference. Further, we show how these measures can be combined and how a priori information obtained in an iterative loop can be incorporated into the proposed scheme. Simulation results incorporating realistic channel and interference models demonstrate the effectiveness of the proposed scheme.

Pareto Models, Top Incomes, and Recent Trends in UK Income Inequality

I determine UK income inequality levels and trends by combining inequality estimates from tax return data (for the 'rich') and household survey data (for the 'non-rich'), taking advantage of the better coverage of top incomes in tax return data (which I demonstrate) and creating income variables in the survey data with the same definitions as in the tax data to enhance comparability. For top income recipients, I estimate inequality and mean income by fitting Pareto models to the tax data, examining specification issues in depth, notably whether to use Pareto I or Pareto II (generalised Pareto) models, and the choice of income threshold above which the Pareto models apply. The preferred specification is a Pareto II model with a threshold set at the 99th or 95th percentile (depending on year). Conclusions about aggregate UK inequality trends since the mid-1990s are robust to the way in which tax data are employed. The Gini coefficient for gross individual income rose by around 7% or 8% between 1996/97 and 2007/08, with most of the increase occurring after 2003/04. The corresponding estimate based wholly on the survey data is around –5%.

A landscape-based cluster analysis using recursive search instead of a threshold parameter

Cluster-based analysis methods in neuroimaging provide control of whole-brain false positive rates without the need to conservatively correct for the number of voxels and the associated false negative results. The current method defines clusters based purely on shapes in the landscape of activation, instead of requiring the choice of a statistical threshold that may strongly affect results. Statistical significance is determined using permutation testing, combining both size and height of activation. A method is proposed for dealing with relatively small local peaks. Simulations confirm the method controls the false positive rate and correctly identifies regions of activation. The method is also illustrated using real data.•A landscape-based method to define clusters in neuroimaging data avoids the need to pre-specify a threshold to define clusters.•The implementation of the method works as expected, based on simulated and real data.•The recursive method used for defining clusters, the method used for combining clusters, and the definition of the “value” of a cluster may be of interest for future variations.

A symmetry-based method to infer structural brain networks from probabilistic tractography data.

A symmetry-based method to infer structural brain networks from probabilistic tractography data.

You Are Probably Not the Weakest Link: Towards Practical Prediction of Susceptibility to Semantic Social Engineering Attacks

Semantic social engineering attacks are a pervasive threat to computer and communication systems. By employing deception rather than by exploiting technical vulnerabilities, spear-phishing, obfuscated URLs, drive-by downloads, spoofed websites, scareware, and other attacks are able to circumvent traditional technical security controls and target the user directly. Our aim is to explore the feasibility of predicting user susceptibility to deception-based attacks through attributes that can be measured, preferably in real-time and in an automated manner. Toward this goal, we have conducted two experiments, the first on 4333 users recruited on the Internet, allowing us to identify useful high-level features through association rule mining, and the second on a smaller group of 315 users, allowing us to study these features in more detail. In both experiments, participants were presented with attack and non-attack exhibits and were tested in terms of their ability to distinguish between the two. Using the data collected, we have determined practical predictors of users’ susceptibility against semantic attacks to produce and evaluate a logistic regression and a random forest prediction model, with the accuracy rates of. 68 and. 71, respectively. We have observed that security training makes a noticeable difference in a user’s ability to detect deception attempts, with one of the most important features being the time since last self-study, while formal security education through lectures appears to be much less useful as a predictor. Other important features were computer literacy, familiarity, and frequency of access to a specific platform. Depending on an organisation’s preferences, the models learned can be configured to minimise false positives or false negatives or maximise accuracy, based on a probability threshold. For both models, a threshold choice of 0.55 would keep both false positives and false negatives below 0.2.

Vectorization Method on the Color Cloud Image

The vectorization, which is an interdisciplinary combining subjects of the computer vision, computer image processing, computer graphics and artificial intelligence, is an important issue in the digital image processing. This paper makes a study of the vectorization method on the satellite color cloud image, Specific techniques such as cloud image layering, threshold selection, edge detection and tracking, feature point extraction, curve fitting are involved. This paper proposes a color layering algorithm whose parameter is the difference between luminance and saturability by analyzing the color features of the satellite cloud pictures. In this algorithm method, we first generate the reference image and then histogram equalize the reference image, afterwards do de-noises process to the above histogram and smooth it for the threshold choice. After analyzing the uncertainty of satellite cloud image edge, this paper puts forward an edge detection algorithm based on adaptive threshold. In order to modify the shortcomings of edge data of raster and edge data of chain code, this paper puts forward the corner detection method based on edge data chain. Also this paper develops a prototype system of automatic vectorization on satellite cloud images by curve fitting the feature points and gets the final vector graphics. Experiments show that the system has the basic functions of automatic extraction, recognition and vectorization for the information of satellite cloud clouds, and carry out the comprehensive verification and analysis for the theory and algorithm by using the real satellite cloud image data.