Likelihood Ration Research Articles

Trypsinogen-2 and trypsinogen activation peptide (TAP) in urine of patients with acute pancreatitis

Background and aims. There is an obvious clinical need for a simple test that can identify patients at risk of developing severe acute pancreatitis. In this work we compared urinary trypsinogen-2 with urinary trypsinogen activation peptide (TAP) and serum C-reactive protein (CRP) for early differentiation between mild and severe acute pancreatitis. Patients and methods. The study population consisted of 127 consecutive patients with acute pancreatitis of whom 29 had severe disease. Urinary trypsinogen-2 was measured by a quantitative immunofluorometric assay and TAP by a competitive immunoassay. Serum CRP was determined by immunoturbidimetry. Results. The sensitivity and specificity to identify severe acute pancreatitis on admission was 72% and 81% for urinary trypsinogen-2, 64% and 82% for urinary TAP, and 29% and 93% for serum CRP, respectively. At 24 h after admission, the values were 82% and 78% for urinary trypsinogen-2, 52% and 92% for urinary TAP, and 84% and 72% for serum CRP, respectively. Receiver-operating characteristics curve analysis showed that the area under the curve was larger for urinary trypsinogen-2 than for urinary TAP and serum CRP on admission and 24 h after admission. On admission the positive likelihood ration for urinary trypsiongen-2 was 3.7, for urinary TAP 3.6, and 4.3 for serum CRP, respectively. The corresponding negative likelihood ratios were 0.34, 0.43, and 0.76, respectively. Conclusion. Urinary trypsinogen-2 was superior to serum CRP and as god as or even better than urinary TAP and in the early prediction of disease severity in acute pancreatitis. These results suggest that it could be a valuable adjunct in the early assessment of the severity of acute pancreatitis.

Minimax estimation of sharp change points

We define the sharp change point problem as an extension of earlier problems in change point analysis related to nonparametric regression. As particular cases, these include estimation of jump points in smooth curves. More generally, we give a systematic treatment of the correct rate of convergence for estimating the position of a “cusp”of an arbitrary order. We propose a test function for the local regularity of a signal that characterizes such a point as a global maximum. In the sample implementation of our method, from observations of the signal at discrete time positions $i/n, i =1 \ldots,n$, we use a wavelet transformation to approximate the position of the change point in the no-noise case. We study the noise effect, in the worst case scenario over a wide class of functions having a unique irregularity of “order $\alpha$” and propose a sequence of estimators which converge at the rate $n_{-1/(1+2\alpha)}$, as $n$ tends to infinity. Finally we analyze the likelihood ration of the problem and show that this is actually the minimaz rate of convertence. Examples of thresholding empirical wavelet coefficients to estimate the position of sharp change points are also presented.

Unsupervised speaker clustering for automatic speaker indexing of recorded audio data

A system and method for unsupervised clustering of audio data segments in an audio data recording containing speech from multiple speakers including the steps of: 1) providing a portion of the audio data containing speech from all of the speakers; 2) forming initial clusters by dividing the portion of the audio data into segments, each of which includes an ordered data set; 3) computing the pairwise distance between each pair of clusters using a likelihood ration independent of the order of data within the segments; and 4) combining into a new cluster the two clusters with a minimum pairwise distance. These steps are repeated until a number of clusters equal to the number of speakers is obtained.

The possible meaning of transferrin and its soluble receptors in seminal plasma as markers of the seminiferous epithelium

Transferrin (Tf) and soluble transferrin receptors (S-Tf-R) were measured by enzyme immunoassay in seminal plasma of 130 semen samples. The mean concentration of S-Tf-R in cases with normozoospermia was 10.4 IU/ml (95% confidence interval: 9.5-11.3) and it was significantly lower in patients with oligozoospermia (6.6, 95% CI: 5.8-7.5, P < 0.001), asthenozoospermia (8.5, 95% CI: 5.5-10.7, P < 0.05), azoospermia of primary testicular origin (7.9, 95% CI: 6.1-9.6, P<0.05) and post-vasectomy samples (5.9, 95% CI: 5.4-6.9, P < 0.001). The concentration of S-Tf-R in post-vasectomy samples was lower than that in patients with azoospermia of primary testicular origin (P < 0.05; positive likelihood ration = 7 at value of 8.3 IU/ ml). S-Tf-R was positively correlated with motile sperm concentration (r = 0.50, P < 0.0001), percentage motility (r = 0.38, P < 0.001), percentage of normal forms (r = 0.43, P < 0.001), sperm linear velocity (r = 0.42, P < 0.001), and ATP concentration (r = 0.67, P < 0.0001). Follicle stimulating hormone (FSH) was found to be negatively correlated with the concentrations of both Tf (r = -0.31, P < 0.05) and of S-Tf-R (r = -0.45, P < 0.01). The mean concentration of Tf in seminal plasma was 50.4 micrograms/ml (35.9-67.2) in samples with normozoospermia (n = 22), and the concentration was significantly lower in patients with oligozoospermia (P < 0.05), azoospermia of testicular origin (P < 0.001), and post-vasectomy samples (P < 0.001). Seminal Tf was correlated with motile sperm concentration (r = 0.36, P < 0.001), percentage of motile spermatozoa (r = 0.25, P < 0.05), linear velocity (r = 0.24, P < 0.05) and ATP concentration (r = p.44, P < 0.001). The concentration of Tf was positively correlated with that of S-Tf-R both in cases with spermatozoa present (r = 0.66, P < 0.001), and in cases with azoospermia of testicular origin (r = 0.51, P < 0.05) but not in vasectomy cases. It is concluded that S-Tf-R in seminal plasma is a marker of spermatogenesis and may give information on the presence or absence of spermatogenetic cells in cases with azoospermia. Further investigations are needed to assess its usefulness for clinical practice.

On the Rate of Approximations for Maximum Likelihood Tests in Change-Point Models

We study the asymptotics of maximum-likelihood ratio-type statistics for testing a sequence of observations for no change in parameters against a possible change while some nuisance parameters remain constant over time. We obtain extreme value as well as Gaussian-type approximations for the likelihood ratio. We get necessary and sufficient conditions for the weak convergence of supremum andLp-functionals of the likelihood ration process. We also approximate the maximum likelihood ratio with Ornstein-Uhlenbeck processes and obtain bounds for the rate of approximation. We show that the Ornstein-Uhlenbeck approach is superior to the extreme value limit in case of moderate sample sizes.

Discrimination of AR, MA and ARMA time series models

The problem of discrimination between two stationary ARMA time series models is considered, and in particular AR(p), MA(p), ARMA(1,1) models. The discriminant based on the likelihood ration leads to a quadratic form that is generally too complicated to evaluated explicitly. The discriminant can be expressed approximately as a linear combination of independent chi–squared random varianles each with one degree of freedom, the coefficients, of which are eigenvalues of cumbersome matrices. An analytical solution which gives the coefficients approximately is suggested.

On the estimation of the stochastic intensity and the parameters of neyman-scott trigger processes

In this paper we are concerned with a class of simple point processes, whose unobservable stochastic intensity is a shot-noise process. We derive a stochastic equation for the conditional moment generating function of the intensity, which can be solved in a recursive way. This yields explicit expression for the minimum variance estimate of the intensity as well as the likelihood ration with respect to the reference measure, on the basis of point process observations.

Asymptotic testing theory for generalized linear models

Statistical inference in generalized linear models is based on the premises that the maximum likelihood estimator of unknown parameters is consistent and asymptotically normal, and that various test statistics have a limiting x2-distribution. FAHRMEIR and KAUFMANN (1985) present mild conditions which assure consistency and asymptotic normality of the maximum likelihood estimator. In this paper it is shown that under essentially tha same conditions the likelihood ration statistic, the Wald statistics and the score statistic are asymptotically equivalent, i.e. they have the same limiting x2-distributions under the general linear hypothesis as well as under suitable sequences of alternatives. Thus, statistical inference in generalized linear models is asymptotically justified under rather weak requirements.

Expectations and the Forward Exchange Rate

This paper provides an empirical examination of the hypothesis that the forward exchange rate provides an optimal forecast of the future spot ex-change rate, for five currencies relative to the dollar. This hypothesis provides a convenient norm for examining the erratic behavior of exchange rates; this erratic behavior represents an efficient market that is quickly incorporating new information into the current exchange rate. This hypothesis is analyzed using two distinct, but related, approaches. The first approach is based on a regression of spot rates on lagged forward rates. When using weekly data and a one month forward exchange rate, ordinary least squares regression analysis of market efficiency is incorrect. Econometric methods are proposed which allow for consistent (though not fully efficient) estimation of the parameters and their standard errors. This paper also presents a new approach for testing exchange market efficiency. This approach is based on a general time series process generating the spot and forward exchange rate. The hypothesis of efficiency implies a set of cross-equation restrictions imposed on the parameters of the time series model. This paper derives these restrictions, proposes a maximum likelihood method of estimating the constrained likelihood function, estimates the model and tests the validity of the restrictions with a likelihood ration statistic.