Profile Likelihood Ratio Research Articles

Limits on Spin-Dependent WIMP-Nucleon Cross Section Obtained from the Complete LUX Exposure.

We present experimental constraints on the spin-dependent WIMP-nucleon elastic cross sections from the total 129.5 kg yr exposure acquired by the Large Underground Xenon experiment (LUX), operating at the Sanford Underground Research Facility in Lead, South Dakota (USA). A profile likelihood ratio analysis allows 90%C.L. upper limits to be set on the WIMP-neutron (WIMP-proton) cross section of σ_{n}=1.6×10^{-41} cm^{2} (σ_{p}=5×10^{-40} cm^{2}) at 35 GeV c^{-2}, almost a sixfold improvement over the previous LUX spin-dependent results. The spin-dependent WIMP-neutron limit is the most sensitive constraint to date.

Results on the Spin-Dependent Scattering of Weakly Interacting Massive Particles on Nucleons from the Run 3 Data of the LUX Experiment.

We present experimental constraints on the spin-dependent WIMP (weakly interacting massive particle)-nucleon elastic cross sections from LUX data acquired in 2013. LUX is a dual-phase xenon time projection chamber operating at the Sanford Underground Research Facility (Lead, South Dakota), which is designed to observe the recoil signature of galactic WIMPs scattering from xenon nuclei. A profile likelihood ratio analysis of 1.4×10^{4} kg day of fiducial exposure allows 90%C.L. upper limits to be set on the WIMP-neutron (WIMP-proton) cross section of σ_{n}=9.4×10^{-41} cm^{2} (σ_{p}=2.9×10^{-39} cm^{2}) at 33 GeV/c^{2}. The spin-dependent WIMP-neutron limit is the most sensitive constraint to date.

Physics reach of the XENON1T dark matter experiment.

The XENON1T experiment is currently in the commissioning phase at the Laboratori Nazionali del Gran Sasso, Italy. In this article we study the experiment's expected sensitivity to the spin-independent WIMP-nucleon interaction cross section, based on Monte Carlo predictions of the electronic and nuclear recoil backgrounds.The total electronic recoil background in 1 tonne fiducial volume and (1, 12) keV electronic recoil equivalent energy region, before applying any selection to discriminate between electronic and nuclear recoils, is (1.80 ± 0.15) · 10−4 (kg·day·keV)−1, mainly due to the decay of 222Rn daughters inside the xenon target. The nuclear recoil background in the corresponding nuclear recoil equivalent energy region (4, 50) keV, is composed of (0.6 ± 0.1) (t·y)−1 from radiogenic neutrons, (1.8 ± 0.3) · 10−2 (t·y)−1 from coherent scattering of neutrinos, and less than 0.01 (t·y)−1 from muon-induced neutrons. The sensitivity of XENON1T is calculated with the Profile Likelihood Ratio method, after converting the deposited energy of electronic and nuclear recoils into the scintillation and ionization signals seen in the detector. We take into account the systematic uncertainties on the photon and electron emission model, and on the estimation of the backgrounds, treated as nuisance parameters. The main contribution comes from the relative scintillation efficiency ℒeff, which affects both the signal from WIMPs and the nuclear recoil backgrounds. After a 2 y measurement in 1 t fiducial volume, the sensitivity reaches a minimum cross section of 1.6 · 10−47 cm2 at mχ = 50 GeV/c2.

Likelihood‐based Inference with Missing Data Under Missing‐at‐Random

AbstractLikelihood‐based inference with missing data is challenging because the observed log likelihood is often an (intractable) integration over the missing data distribution, which also depends on the unknown parameter. Approximating the integral by Monte Carlo sampling does not necessarily lead to a valid likelihood over the entire parameter space because the Monte Carlo samples are generated from a distribution with a fixed parameter value.We consider approximating the observed log likelihood based on importance sampling. In the proposed method, the dependency of the integral on the parameter is properly reflected through fractional weights. We discuss constructing a confidence interval using the profile likelihood ratio test. A Newton–Raphson algorithm is employed to find the interval end points. Two limited simulation studies show the advantage of the Wilks inference over the Wald inference in terms of power, parameter space conformity and computational efficiency. A real data example on salamander mating shows that our method also works well with high‐dimensional missing data.

Directional detection of dark matter streams

Directional detection of WIMPs, in which the energies and directions of the recoiling nuclei are measured, currently presents the only prospect for probing the local velocity distribution of Galactic dark matter. We investigate the extent to which future directional detectors would be capable of probing dark matter substructure in the form of streams. We analyse the signal expected from a Sagittarius-like stream and also explore the full parameter space of stream speed, direction, dispersion and density. Using a combination of non-parametric directional statistics, a profile likelihood ratio test and Bayesian parameter inference we find that within acceptable exposure times (O(10) kg yr for cross sections just below the current exclusion limits) future directional detectors will be sensitive to a wide range of stream velocities and densities. We also examine and discuss the importance of the energy window of the detector.

Higgs production in e-e+ collisions as a probe of noncommutativity

We examine the sensitivity of the angular distribution of the Higgs boson in the process of $e^+e^-\to Z H$ and the total cross section in the minimal noncommutative standard model (mNCSM) framework to set lower limit on the noncommutative charactristic scale ($\Lambda$). Contrary to the standard model case, in this process the Higgs boson tends to be emitted anisotropically in the transverse plane. Based on this fact, the profile likelihood ratio is used to set lower limit on $\Lambda$. The lower limit is presented as a function of the integrated luminosity. We show that at the center-of-mass energy of 1.5 TeV and with 500 fb$^{-1}$ of data, the noncommutative characteristic energy scale $\Lambda$ can be excluded up to 1.2 TeV.

Silicon Detector Dark Matter Results from the Final Exposure of CDMS II

We report results of a search for weakly interacting massive particles (WIMPS) with the silicon detectors of the CDMS II experiment. This blind analysis of 140.2 kg day of data taken between July 2007 and September 2008 revealed three WIMP-candidate events with a surface-event background estimate of 0.41(-0.08)(+0.20)(stat)(-0.24)(+0.28)(syst). Other known backgrounds from neutrons and 206Pb are limited to <0.13 and <0.08 events at the 90% confidence level, respectively. The exposure of this analysis is equivalent to 23.4 kg day for a recoil energy range of 7-100 keV for a WIMP of mass 10 GeV/c2. The probability that the known backgrounds would produce three or more events in the signal region is 5.4%. A profile likelihood ratio test of the three events that includes the measured recoil energies gives a 0.19% probability for the known-background-only hypothesis when tested against the alternative WIMP+background hypothesis. The highest likelihood occurs for a WIMP mass of 8.6 GeV/c2 and WIMP-nucleon cross section of 1.9×10(-41) cm2.

Profile Likelihood Inferences on the Partially Linear Model with a Diverging Number of Parameters

In this article, we study the profile likelihood estimation and inference on the partially linear model with a diverging number of parameters. Polynomial splines are applied to estimate the nonparametric component and we focus on constructing profile likelihood ratio statistic to examine the testing problem for the parametric component in the partially linear model. Under some regularity conditions, the asymptotic distribution of profile likelihood ratio statistic is proposed when the number of parameters grows with the sample size. Numerical studies confirm our theory.

Alternatives to the usual likelihood ratio test in mixed linear models

The small-sample performance of alternatives to the usual likelihood ratio test in mixed linear models is investigated. Specifically, the following tests for fixed effects are considered: (i) a bootstrap-based test, (ii) the Bartlett-corrected usual test, and (iii) an adjusted profile likelihood ratio test. The last test is derived using an approximation to the modified profile likelihood proposed by Barndorff-Nielsen, based on the work of Severini. Bootstrap resampling is performed to numerically construct a Bartlett correction factor for the usual test statistic, and also to obtain a critical value that does not rely on first-order asymptotics. The numerical evidence presented in the paper slightly favors the Bartlett-corrected usual test. An application to real longitudinal data is presented.

Semiparametric inference with a functional-form empirical likelihood

A functional-form empirical likelihood method is proposed as an alternative method to the empirical likelihood method. The proposed method has the same asymptotic properties as the empirical likelihood method but has more flexibility in choosing the weight construction. Because it enjoys the likelihood-based interpretation, the profile likelihood ratio test can easily be constructed with a chi-square limiting distribution. Some computational details are also discussed, and results from finite-sample simulation studies are presented.

On Consistencies of the Profile Likelihood Estimators and their Derivatives of the Distribution Function in Doubly Censored Data

We consider the semiparametric profile likelihood inference for the distribution function under doubly censored data. For further developments of the statistical inference based on the profile likelihood ratio and alternative tools such as the score or Wald-type inference, we discuss the structures of the profile likelihood estimators and their derivatives included in the score function and the Fisher function of the profile likelihood, establishing the consistencies of their estimators.

Statistical inference for partially time-varying coefficient errors-in-variables models

This paper studies the partially time-varying coefficient models where some covariates are measured with additive errors. In order to overcome the bias of the usual profile least squares estimation when measurement errors are ignored, we propose a modified profile least squares estimator of the regression parameter and construct estimators of the nonlinear coefficient function and error variance. The proposed three estimators are proved to be asymptotically normal under mild conditions. In addition, we introduce the profile likelihood ratio test and then demonstrate that it follows an asymptotically χ2 distribution under the null hypothesis. Finite sample behavior of the estimators is investigated via simulations too.

Improved Birnbaum–Saunders inference under type II censoring

The Birnbaum–Saunders distribution is useful for modeling reliability data. In this paper we obtain adjusted profile maximum likelihood estimators for the Birnbaum–Saunders distribution shape parameter under type II data censoring. We consider the adjustments to the profile likelihood function proposed by Barndorff-Nielsen (1983), Cox and Reid (1987), Fraser and Reid (1995), Fraser et al. (1999) and Severini (1998, 1999). We obtain modified profile likelihood ratio tests and also consider bootstrap-based inference. Interval estimation is addressed as well. Several bootstrap confidence intervals are considered. Monte Carlo simulation results on point estimation, interval estimation and hypothesis testing inference are reported. Finally, we present two applications that use real (not simulated) data.

Exploring the varying covariate effects in proportional odds models with censored data

In this article, we consider a proportional odds model, which allows one to examine the extent to which covariates interact nonlinearly with an exposure variable for analysis of right-censored data. A local maximum likelihood approach is presented to estimate nonlinear interactions (the coefficient functions) and the baseline function. The proposed estimators are shown to be consistent and asymptotically normal with the asymptotic variance estimated consistently. Also, we develop local profile likelihood ratio method to construct confidence region of coefficient functions. Simulation studies are conducted to evaluate the performances of the proposed estimators, and compare the normal approximation based confidence regions and local profile likelihood ratio based confidence regions. The method is illustrated with Stanford heart transplant data.

Assessing the discovery potential of directional detection of dark matter

There is a worldwide effort toward the development of a large TPC (Time Projection Chamber) devoted to directional Dark Matter detection. All current projects are being designed to fulfill a unique goal : identifying weakly interacting massive particle (WIMP) as such by taking advantage of the expected direction dependence of WIMP-induced events toward the constellation Cygnus. However such proof of discovery requires a careful statistical data treatment. In this paper, the discovery potential of forthcoming directional detectors is adressed by using a frequentist approach based on the profile likelihood ratio test statistic. This allows us to estimate the expected significance of a Dark Matter detection. Moreover, using this powerful test statistic, it is possible to propagate astrophysical and experimental uncertainties in the determination of the discovery potential of a given directional detection experiment. This way, we found that a 30 kg.year CF$_4$ directional experiment could reach a 3$\sigma$ sensitivity at 90% C.L. down to $10^{-5}$ pb and $3.10^{-4}$ pb for the WIMP-proton axial cross section in the most optimistic and pessimistic scenario respectively.

The profile likelihood ratio and the look elsewhere effect in high energy physics

The experimental issue of the search for new particles of unknown mass poses the challenge of exploring a wide interval to look for the usual signatures represented by excess of events above the background. A side effect of such a broad range quest is that the significance calculations valid for signals of known location are no more applicable when such an information is missing. This circumstance is commonly termed in high energy physics applications as the look elsewhere effect. How it concretely manifests in a specific problem of signal search depends upon the particular strategy adopted to unravel the sought-after signal from the underlying background. In this respect an increasingly popular method is the profile likelihood ratio, especially because of its asymptotic behavior dictated by one of the most famous statistic result, the Wilks' theorem. This work is centered on the description of the look elsewhere effect in the framework of the profile likelihood methodology, in particular proposing a conjecture about the distribution of the likelihood ratio under the null hypothesis of absence of the signal over the searched mass interval, a condition which is known to violate the conditions of the Wilks' theorem. Several Monte Carlo tests to support the conjecture are presented, as well.

Likelihood approach to the first dark matter results from XENON100

Many experiments that aim at the direct detection of Dark Matter are able to distinguish a dominant background from the expected feeble signals, based on some measured discrimination parameter. We develop a statistical model for such experiments using the Profile Likelihood ratio as a test statistic in a frequentist approach. We take data from calibrations as control measurements for signal and background, and the method allows the inclusion of data from Monte Carlo simulations. Systematic detector uncertainties, such as uncertainties in the energy scale, as well as astrophysical uncertainties, are included in the model. The statistical model can be used to either set an exclusion limit or to make a discovery claim, and the results are derived with a proper treatment of statistical and systematic uncertainties. We apply the model to the first data release of the XENON100 experiment, which allows to extract additional information from the data, and place stronger limits on the spin-independent elastic WIMP-nucleon scattering cross-section. In particular, we derive a single limit, including all relevant systematic uncertainties, with a minimum of 2.4x10^-44 cm^2 for WIMPs with a mass of 50 GeV/c^2.

Challenges of profile likelihood evaluation in multi-dimensional SUSY scans

Statistical inference of the fundamental parameters of supersymmetric theories is a challenging and active endeavor. Several sophisticated algorithms have been employed to this end. While Markov-Chain Monte Carlo (MCMC) and nested sampling techniques are geared towards Bayesian inference, they have also been used to estimate frequentist conifidence intervals based on the profile likelihood ratio. We investigate the performance and appropriate configuration of MultiNest, a nested sampling based algorithm, when used for profile likelihood-based analyses both on toy models and on the parameter space of the Constrained MSSM. We find that while the standard configuration previously used in the literarture is appropriate for an accurate reconstruction of the Bayesian posterior, the profile likelihood is poorly approximated. We identify a more appropriate MultiNest configuration for profile likelihood analyses, which gives an excellent exploration of the profile likelihood (albeit at a larger computational cost), including the identification of the global maximum likelihood value. We conclude that with the appropriate configuration MultiNest is a suitable tool for profile likelihood studies, indicating previous claims to the contrary are not well founded.

Erratum to: A Wald-type variance estimation for the nonparametric distribution estimators for doubly censored data

We discuss the variance estimation for the nonparametric distribution estimator for doubly censored data. We first provide another view of Kuhn–Tucker’s conditions to construct the profile likelihood, and lead a Newton–Raphson algorithm as an optimization technique unlike the EM algorithm. The main proposal is an iteration-free Wald-type variance estimate based on the chain rule of differentiating conditions to construct the profile likelihood, which generalizes the variance formula in only right- or left-censored data. In this estimation procedure, we overcome some difficulties caused in directly applying Turnbull’s formula to large samples and avoid a load with computationally heavy iterations, such as solving the Fredholm equations, computing the profile likelihood ratio or using the bootstrap. Also, we establish the consistency of the formulated Wald-type variance estimator. In addition, simulation studies are performed to investigate the properties of the Wald-type variance estimates in finite samples in comparison with those from the profile likelihood ratio.

Likelihood Methods for Regression Models with Expensive Variables Missing by Design

In some applications involving regression the values of certain variables are missing by design for some individuals. For example, in two-stage studies (Zhao and Lipsitz, 1992), data on "cheaper" variables are collected on a random sample of individuals in stage I, and then "expensive" variables are measured for a subsample of these in stage II. So the "expensive" variables are missing by design at stage I. Both estimating function and likelihood methods have been proposed for cases where either covariates or responses are missing. We extend the semiparametric maximum likelihood (SPML) method for missing covariate problems (e.g. Chen, 2004; Ibrahim et al., 2005; Zhang and Rockette, 2005, 2007) to deal with more general cases where covariates and/or responses are missing by design, and show that profile likelihood ratio tests and interval estimation are easily implemented. Simulation studies are provided to examine the performance of the likelihood methods and to compare their efficiencies with estimating function methods for problems involving (a) a missing covariate and (b) a missing response variable. We illustrate the ease of implementation of SPML and demonstrate its high efficiency.