Density Histograms Research Articles

An Improved Test for Earnings Management Using Kernel Density Estimation

This paper describes improvements on methods developed by Burgstahler and Dichev (1997, Earnings management to avoid earnings decreases and losses, Journal of Accounting and Economics, 24(1), pp. 99–126) and Bollen and Pool (2009, Do hedge fund managers misreport returns? Evidence from the pooled distribution, Journal of Finance, 64(5), pp. 2257–2288) to test for earnings management by identifying discontinuities in distributions of scaled earnings or earnings forecast errors. While existing methods use preselected bandwidths for kernel density estimation and histogram construction, the proposed test procedure addresses the key problem of bandwidth selection by using a bootstrap test to endogenise the selection step. The main advantage offered by the bootstrap procedure over prior methods is that it provides a reference distribution that cannot be globally distinguished from the empirical distribution rather than assuming a correct reference distribution. This procedure limits the researcher's degrees of freedom and offers a simple procedure to find and test a local discontinuity. I apply the bootstrap density estimation to earnings, earnings changes, and earnings forecast errors in US firms over the period 1976–2010. Significance levels found in earlier studies are greatly reduced, often to insignificant values. Discontinuities cannot be detected in analysts’ forecast errors, while such findings of discontinuities in earlier research can be explained by a simple rounding mechanism. Earnings data show a large drop in loss aversion after 2003 that cannot be detected in changes of earnings.

Repeatability and Sample Size Assessment Associated with Computed Tomography-Based Lung Density Metrics.

Density-based metrics assess severity of lung disease but vary with lung inflation and method of scanning. The aim of this study was to evaluate the repeatability of single center, CT-based density metrics of the lung in a normal population and assess study sample sizes needed to detect meaningful changes in lung density metrics when scan parameters and volumes are tightly controlled. Thirty-seven subjects (normal smokers and non-smokers) gave consent to have randomly assigned repeated, breath-held scans at either inspiration (90% vital capacity: TLC) or expiration (20% vital capacity: FRC). Repeated scans were analyzed for: mean lung density (MLD), 15th percentile point of the density histogram (P15), low attenuation areas (LAA) and alpha (fractal measure of hole size distribution). Using inter-subject differences and previously reported bias, sample size was estimated from month or yearly change in density metrics obtained from published literature (i.e. meaningful change). Inter-scan difference measurements were small for density metrics (ICC > 0.80) and average ICCs for whole lung alpha-910 and alpha-950 were 0.57 and 0.64, respectively. Power analyses demonstrated that, under the control conditions with minimal extrinsic variation, population sizes needed to detect meaningful changes in density measures for TLC or FRC repeated scans ranged from a few (20-40) to a few hundred subjects, respectively. A meaningful sample size was predicted from this study using volume-controlled normal subjects in a controlled imaging environment. Under proper breath-hold conditions, high repeatability was obtained in cohorts of normal smokers and non-smokers.

Quantum effects and anharmonicity in the H2-Li+-benzene complex: A model for hydrogen storage materials

Quantum and anharmonic effects are investigated in H2-Li(+)-benzene, a model for hydrogen adsorption in metal-organic frameworks and carbon-based materials. Three- and 8-dimensional quantum diffusion Monte Carlo (QDMC) and rigid-body diffusion Monte Carlo (RBDMC) simulations are performed on potential energy surfaces interpolated from electronic structure calculations at the M05-2X/6-31+G(d,p) and M05-2X/6-311+G(2df,p) levels of theory using a three-dimensional spline or a modified Shepard interpolation. These calculations investigate the intermolecular interactions in this system, with three- and 8-dimensional 0 K H2 binding enthalpy estimates, ΔH(bind) (0 K), being 16.5 kJ mol(-1) and 12.4 kJ mol(-1), respectively: 0.1 and 0.6 kJ mol(-1) higher than harmonic values. Zero-point energy effects are 35% of the value of ΔH(bind) (0 K) at M05-2X/6-311+G(2df,p) and cannot be neglected; uncorrected electronic binding energies overestimate ΔHbind (0 K) by at least 6 kJ mol(-1). Harmonic intermolecular binding enthalpies can be corrected by treating the H2 "helicopter" and "ferris wheel" rotations as free and hindered rotations, respectively. These simple corrections yield results within 2% of the 8-dimensional anharmonic calculations. Nuclear ground state probability density histograms obtained from the QDMC and RBDMC simulations indicate the H2 molecule is delocalized above the Li(+)-benzene system at 0 K.

Quantitative kinetic inhibitor comparisons and memory effect measurements from hydrate formation probability distributions

In contrast with conventional experiments for ranking the performance of kinetic hydrate inhibitors (KHIs), the high pressure automated lag time apparatus (HP-ALTA) is not limited by the stochasticity of hydrate formation because it can make large numbers of formation measurements rapidly. An improved procedure for analysing HP-ALTA data is presented here in which formation probability density histograms with uniform temperature bin widths are constructed. Operations such as numerical integration and subtraction can then be applied to entire formation distributions without the need to regress the data to model analytic functions. This approach enables apparatus effects on the measured formation distributions to be quantified and mitigated. Using analogue natural gas mixtures, a demonstration of the improved method is reported here for six KHIs that were also tested using a conventional rocking cell. The cumulative formation probability distribution functions determined by numerical integration of the histograms were used to quantitatively (1) assess the performance of the KHIs, (2) study the memory effect in pure water, and (3) assess its impact on the performance of each KHI. The rankings of the KHIs obtained from the various performance metrics accessible through the HP-ALTA data were generally consistent with the rocking cell rankings, with the average absolute deviation in net subcooling for the two apparatus being 2 to 3K. In the quiescent HP-ALTA samples, the three worst performing KHIs still had expected subcoolings of about 14K unless the memory effect was present, in which case the expected net subcoolings of the three worst performing KHIs ranged from 0–6K. In contrast the memory effect reduced the expected subcoolings of the three best performing KHIs by only 1–2K. Using the new method of analysis, hydrate formation distributions measured with the HP-ALTA can also be used to identify regions of good and poor performance for a given KHI, which might aid in the determination of improved inhibition strategies or assist the development of improved hydrate inhibitors.

Kurtosis and skewness assessments of solid lung nodule density histograms: differentiating malignant from benign nodules on CT

The purpose of our study was to assess pulmonary nodule characteristics using density histogram kurtosis and skewness and to distinguish malignant from benign nodules. Ninety-three lung nodules on CT were analyzed, including 72 malignant and 21 benign nodules. They were completely solid or solid with limited ground-glass opacity. Based on their CT characteristics, nodules were categorized into type A, homogeneous nodules with uniform internal structures and clear margins, and type B, inhomogeneous nodules with heterogeneous structures or uneven margins. Kurtosis and skewness were calculated from density histograms to compare type A and B nodules and malignant and benign nodules. Receiver-operating characteristic (ROC) curves were generated to assess kurtosis and skewness for discriminating between different nodule types. Type A nodules (n = 35) had greater kurtosis and reduced skewness (p < 0.001) compared to type B nodules (n = 58). Malignant tumor kurtosis was greater than that of benign nodules (type A, p < 0.05; type B, p = 0.001). Type B malignant tumors had reduced skewness compared to benign nodules (p < 0.05). ROC curves provided relatively high values for the area under the curve (0.71-0.83). Kurtosis and skewness assessments of density histograms may be useful for differentiating malignant from benign nodules.

Spatial forest structure reconstruction as a strategy for mitigating edge-bias in circular monitoring plots

Forest ecosystem monitoring is an important task of forest science worldwide and its outcomes have important consequences both for national forest policies and local decision making. Often forest monitoring is implemented using circular plots of limited size. Edge effects can seriously bias the results of spatially explicit analyses of circular plot data and little research has been carried out on how to mitigate this problem. In this study, we have compared the method of spatial forest structure reconstruction to traditional plus-sampling, to the reflection method and to a situation where no edge-bias mitigation method is used at all. Reconstruction is a non-parametric modelling method based on simulated annealing. In the context of this study, the arithmetic means of structural summary characteristics are used to extrapolate spatial patterns previously measured in the core area of the monitoring plots to the margins outside. The computer experiments were based on 706 circular monitoring plots of the Estonian long-term monitoring network maintained by the Estonian University of Life Sciences, Tartu. The results clearly indicate the superiority of the reconstruction method and suggest that this approach has great potential for future spatially-explicit data analyses and modelling involving circular monitoring plots. Further improvements can be expected from using density functions and histograms of structural summary characteristics instead of arithmetic means.

Semi-automated scoring of pulmonary emphysema from X-ray CT: Trainee reproducibility and accuracy

ObjectiveWe developed a semi-automated tool to quantify emphysema from thoracic X-ray multi-detector (64-slice) computed tomography (CT) for training purposes and multi-reader studies. Materials and MethodsThoracic X-ray CT was acquired in 93 ex-smokers, who were evaluated by six trainees with little or no expertise (trainees) and a single experienced thoracic radiologist (expert). A graphic user interface (GUI) was developed for emphysema quantification based on the percentile of lung where a score of 0=no abnormalities, 1=1–25%, 2=26–50%, 3=51–75% and 4=76–100% for each lung side/slice. Trainees blinded to subject characteristics scored randomized images twice; accuracy was determined by comparison to expert scores, density histogram 15th percentile (HU15), relative area at −950HU (RA950), low attenuation clusters at −950HU (LAC950), −856HU (LAC856) and the diffusing capacity for carbon monoxide (DLCO%pred). Intra- and inter-observer reproducibility was evaluated using coefficients-of-variation (COV), intra-class (ICC) and Pearson correlations. ResultsTrainee–expert correlations were significant (r=0.85–0.97, p<0.0001) and a significant trainee bias (0.15±0.22) was observed. Emphysema score was correlated with RA950 (r=0.88, p<0.0001), HU15 (r=−0.77, p<0.0001), LAC950 (r=0.76, p<0.0001), LAC856 (r=0.74, p=0.0001) and DLCO%pred (r=−0.71, p<0.0001). Intra-observer reproducibility (COV=4–27%; ICC=0.75–0.94) was moderate to high for trainees; intra- and inter-observer COV were negatively and non-linearly correlated with emphysema score. ConclusionWe developed a GUI for rapid and interactive emphysema scoring that allows for comparison of multiple readers with clinical and radiological standards.

A Histogram-Based Segmentation Method for Wideband Spectrum Sensing in Cognitive Radios

This paper presents a new method for automatic radio spectrum segmentation. Spectrum segmentation serves as a first step to identify the subbands that are in use at a certain time, when a wide portion of the spectrum is observed. The proposed method, based on the evaluation of the histogram of the power spectral density of the received signal, is designed for obtaining a low processing complexity and to be implemented in a measurement instrument or a cognitive radio receiver. The proposed method is validated on emulated signals, and the results show good accuracy of the subband boundary estimation and a processing complexity much lower than the other segmentation methods.

Computed Tomography Density Histogram Analysis to Evaluate Pulmonary Emphysema in Ex-smokers

High-resolution computed tomography (CT) measurements of emphysema typically use Hounsfield unit (HU) density histogram thresholds or observer scores based on regions of low x-ray attenuation. Our objective was to develop an automated measurement of emphysema using principal component analysis (PCA) of the CT density histogram. Ninety-seven ex-smokers, including 53 subjects with chronic obstructive pulmonary disease (COPD) and 44 asymptomatic subjects (AEs), provided written informed consent to imaging as well as plethysmography and spirometry. We applied PCA to the CT density histogram to generate whole lung and regional density histogram principal components including the first and second components and the sum of both principal components (density histogram principal component score [DHPCS]). Significant relationships for DHPCS with single HU thresholds, pulmonary function measurements, an expert's emphysema score, and hyperpolarized (3)He magnetic resonance imaging apparent diffusion coefficients (ADCs) were determined using linear regression and Pearson coefficients. Receiver operator characteristics analysis was performed using forced expiratory volume in 1 second (FEV1)/forced vital capacity (FVC) as the independent diagnostic. There was a significant difference (P < .0001) between AE and COPD subjects for DHPCS; FEV1/FVC; diffusing capacity of lung for carbon monoxide%predicted; attenuation values below -950, -910, and -856 HU; and (3)He ADCs. There were significant correlations for DHPCS with FEV1/FVC (r = -0.85, P < .0001); diffusing capacity of lung for carbon monoxide%predicted (r = -0.67, P < .0001); attenuation values below -950/-910/-856 HU (r = 0.93/0.96/0.76, P < .0001); and (3)He ADCs (r = 0.85, P < .0001). Receiver operator characteristics analysis showed a 91% classification rate for DHPCS. We generated an automated emphysema score using PCA of the CT density histogram with a 91% COPD classification rate that showed strong and significant correlations with pulmonary function tests, single HU thresholds, and (3)He magnetic resonance imaging ADCs.

Early craniofacial bone growth and maturation of Fgfr2+/P253R mice and littermates

Our goal was to quantify the maturation of individual craniofacial bones and determine how the Fgfr2 P253R mutation modifies cranial bone maturation. Bone volumes and relative bone density histograms were measured from computed tomography images of Fgfr2+/P253R Apert craniosynostosis syndrome model mice and unaffected littermates at E15.5, E16.5, E17.5, P0, and P2. This data was used to quantify ontogenetic change in individual bone density and to test for differences in relative density between the two genotypes. Individual bones increase in volume and density at different rates across embryonic and perinatal development. While many bones of the Fgfr2+/P253R mice display larger volumes than those of littermates by P0, they generally display smaller volumes at P2. Relative bone density of the frontal and some facial/palatal bones is reduced for Fgfr2+/P253R mice at P2, but not before. These differences suggest a postnatal shift in the effect of the mutation on bone growth and maturation. Some of the most significant differences in craniofacial maturation between the two genotypes occur within the facial skeleton, identified as a region of significant dysmorphology in these mice and in humans carrying mutations for FGFR‐related craniosynostosis syndromes. Work supported in part by grants from the NSF to CJP (BCS‐1061554); the NIDCR/NIH and ARRA to JTR (R01DE018500; 3R01DE018500–02S1; R01DE022988).Grant Funding Source: NIDCR/NIH; NSF

SVM Training Phase Reduction Using Dataset Feature Filtering for Malware Detection

N-gram analysis is an approach that investigates the structure of a program using bytes, characters, or text strings. A key issue with N-gram analysis is feature selection amidst the explosion of features that occurs when N is increased. The experiments within this paper represent programs as operational code (opcode) density histograms gained through dynamic analysis. A support vector machine is used to create a reference model, which is used to evaluate two methods of feature reduction, which are “area of intersect” and “subspace analysis using eigenvectors.” The findings show that the relationships between features are complex and simple statistics filtering approaches do not provide a viable approach. However, eigenvector subspace analysis produces a suitable filter.

Pulmonary ventilation visualized using hyperpolarized helium-3 and xenon-129 magnetic resonance imaging: differences in COPD and relationship to emphysema

In subjects with chronic obstructive pulmonary disease (COPD), hyperpolarized xenon-129 ((129)Xe) magnetic resonance imaging (MRI) reveals significantly greater ventilation defects than hyperpolarized helium-3 ((3)He) MRI. The physiological and/or morphological determinants of ventilation defects and the differences observed between hyperpolarized (3)He and (129)Xe MRI are not yet understood. Here we aimed to determine the structural basis for the differences in ventilation observed between (3)He and (129)Xe MRI in subjects with COPD using apparent diffusion coefficients (ADC) and computed tomography (CT). Ten COPD ex-smokers provided written, informed consent and underwent MRI, CT, spirometry, and plethysmography. (3)He and (129)Xe MRI ventilation volume was generated using semiautomated segmentation, and ADC maps were registered to generate ADC values for lung regions of interest ventilated by both gases (ADCHX) and by (3)He gas only (ADCHO). CT wall area percentage and the lowest 15th percentile point of the CT lung density histogram (HU15%) were also evaluated. For lung regions accessed by (3)He gas only, mean (3)He ADCHO was significantly greater than for regions accessed by both gases (ADCHO = 0.503 ± 0.119 cm(2)/s, ADCHX = 0.470 ± 0.125 cm(2)/s, P < 0.0001). The difference between (3)He and (129)Xe ventilation volume was significantly correlated with CT HU15% (r = -65, P = 0.04) and (3)He ADCHO (r = 0.70, P = 0.02), but not CT wall area percentage (r = -0.34, P = 0.33). In conclusion, in this small study in COPD subjects, we observed significantly decreased (129)Xe MRI ventilation compared with (3)He MRI, and these regions of decreased (129)Xe ventilation were spatially and significantly correlated with regions of increased pulmonary emphysema, but not airway wall thickness.

Automated adjustment of display conditions in brain MR images: diffusion-weighted MRIs and apparent diffusion coefficient maps for hyperacute ischemic stroke

We developed a new computerized scheme for proper display of brain diffusion-weighted magnetic resonance images (DWIs) and apparent diffusion coefficient (ADC) maps based on density histogram analysis. In our scheme, DWI volumes and b0 image volumes of 44 cases were first created, and brain regions on DWI volumes were segmented. ADC map volumes were then derived from both volumes. The density histogram was determined from the brain regions on these volumes, and the voxel value corresponding to the maximum in the density histogram was determined for each volume. The display gray level for each of the two volumes was adjusted by setting the determined voxel value as window conditions. In a comparison between the existing manual method and our automated method, the variation in the gray levels was evaluated quantitatively. The variation in the average of the cross-correlation values determined for pairs of density histograms in each of the DWIs and ADC maps was 57.3 and 27.1 % with the existing method, respectively, and 7.7 and 2.7 % with our scheme, respectively, which indicated a more consistent display of images with our scheme. The performance of the two-alternative-forced-choice method for visual comparison of pairs of images in each of the DWIs and ADC maps adjusted by our scheme was judged to be better than those of the existing method by 75.1 % and 92.7 %, respectively. Our computerized scheme would be a promising technique for an accurate, prompt automated adjustment of display conditions in brain DWIs and ADC maps.

Optical path-length matrix method for estimating skin spectrum

In this article, we propose a new method—the optical path-length matrix method (OPLM)—as a faster alternative to the Monte Carlo for multi-layered media (MCML), which is often used to simulate the skin spectrum. Theoretically, peripheral oxygen saturation can be estimated by iterating MCML, but it is not a realistic strategy because it requires huge computation time. The optical path-length matrix is obtained as the probabilistic density histograms of the optical path length in skin using MCML, and once the matrix is obtained, skin spectral reflectance can be calculated by accumulating all combinations of elements in the matrix and by setting an absorption coefficient based on the Beer-Lambert law. The computational time of OPLM was approximately 26,000 times faster than that of MCML.

Magnetohydrodynamic stability of broad line region clouds

Hydrodynamic stability has been a longstanding issue for the cloud model of the broad line region in active galactic nuclei. We argue that the clouds may be gravitationally bound to the supermassive black hole. If true, stabilisation by thermal pressure alone becomes even more difficult. We further argue that if magnetic fields should be present in such clouds at a level that could affect the stability properties, they need to be strong enough to compete with the radiation pressure on the cloud. This would imply magnetic field values of a few Gauss for a sample of Active Galactic Nuclei we draw from the literature. We then investigate the effect of several magnetic configurations on cloud stability in axi-symmetric magnetohydrodynamic simulations. For a purely azimuthal magnetic field which provides the dominant pressure support, the cloud first gets compressed by the opposing radiative and gravitational forces. The pressure inside the cloud then increases, and it expands vertically. Kelvin-Helmholtz and column density instability lead to a filamentary fragmentation of the cloud. This radiative dispersion continues until the cloud is shredded down to the resolution level. For a helical magnetic field configuration, a much more stable cloud core survives with a stationary density histogram which takes the form of a power law. Our simulated clouds develop sub-Alfvenic internal motions on the level of a few hundred km/s.

Geometrically necessary dislocation density distributions in cyclically deformed Ti–6Al–4V

Cross-correlation-based electron backscattering diffraction analysis has been used to map lattice rotations in Ti–6Al–4V polycrystals deformed by load-controlled fatigue and dwell fatigue including a hold at maximum load. The lattice curvatures were used to form lower bound estimates of the geometrically necessary dislocation (GND) density distributions. In all cases the density of 〈a〉-type GNDs was much higher than for 〈c+a〉-type GNDs. As for tensile deformation the GND density histograms were significantly skewed toward the high density side. Observations of interrupted fatigue tests suggested that the density of 〈a〉-type GNDs decreases during continued cyclic loading, presumably due to the formation of tightly bound dipoles and multipole structures. As has been proposed in models of facet fatigue formation, an example is presented of the accumulation, within a soft grain, of GNDs in a diffuse pile-up against a grain boundary with a hard grain.

A numerical coarse-grained description of a binary alloy

We employ Monte Carlo simulation in the semi-grand canonical ensemble to obtain the coarse-grained free energy corresponding to an embedded-atom method description of a binary alloy. In particular, the Ginzburg-Landau free energy for a Cu-Ni alloy was determined from a tabulated histogram of the joint probability density of composition, energy, and volume. Using histogram reweighting techniques, the free energy is extrapolated to a range of points in parameter space from a small number of simulations. The results are interpreted by comparing the free energy with that corresponding to a regular solution model of an alloy. In addition, we obtain expressions for thermodynamic quantities in terms of the joint cumulants of the probability density at a given temperature and chemical potential difference. These expressions may then be likewise extrapolated to obtain the dependence of the composition on the temperature and the chemical potential difference over a wide range of parameter space.

Gas density histograms of galaxies: the observational density probability function of the interstellar gas density

AbstractIn the steady state, the probability density function (PDF) of the gaseous interstellar matter (ISM) can be observed as a gas density histogram (GDH) of all cells in the system. We made GDHs of the Milky Way Galaxy (MWG) using Galactic plane surveys in CO lines. We found that the GDH in the MWG is log-normal which suggests that the density structure of the molecular gas is a result of many stochastic processes. Using the Nobeyama CO atlas, we made GDHs of nearby galaxies but in column density. Although some galaxies show log-normal, the others show completely different shapes, suggesting that the density structure of galaxies may be different from galaxy to galaxy.

An Efficient Approach for Incremental Association Rule Mining through Histogram Matching Technique

The objective of the work being presented is to propose an approach for obtaining appropriate association rules when the data set is being incrementally updated. During this process raw data is clustered by K-mean Clustering Algorithm and appropriate rules are generated for each cluster. Further, a histogram and probability density function are also generated for each cluster. When Burst data set is coming to the system, initially the histogram and probability density function of this new data set are obtained. The new data set has to be added to the cluster whose histogram and probability density functions are almost similar. The proposed method is evaluated and explained on synthetic data.

MEAN-FIELD CRITICAL BEHAVIOR AND ERGODICITY BREAK IN A NONEQUILIBRIUM ONE-DIMENSIONAL RSOS GROWTH MODEL

We investigate the nonequilibrium roughening transition of a one-dimensional restricted solid-on-solid model by directly sampling the stationary probability density of a suitable order parameter as the surface adsorption rate varies. The shapes of the probability density histograms suggest a typical Ginzburg–Landau scenario for the phase transition of the model, and estimates of the "magnetic" exponent seem to confirm its mean-field critical behavior. We also found that the flipping times between the metastable phases of the model scale exponentially with the system size, signaling the breaking of ergodicity in the thermodynamic limit. Incidentally, we discovered that a closely related model not considered before also displays a phase transition with the same critical behavior as the original model. Our results support the usefulness of off-critical histogram techniques in the investigation of nonequilibrium phase transitions. We also briefly discuss in the appendix a good and simple pseudo-random number generator used in our simulations.