The Experts below are selected from a list of 303 Experts worldwide ranked by ideXlab platform

Thomas S Richardson - One of the best experts on this subject based on the ideXlab platform.

  • an algorithm to compute the Likelihood Ratio Test statistic of the sharp null hypothesis for compliers
    ACI'15 Proceedings of the UAI 2015 Conference on Advances in Causal Inference - Volume 1504, 2015
    Co-Authors: Wen Wei Loh, Thomas S Richardson
    Abstract:

    In a randomized experiment with noncompliance, scientific interest is often in Testing whether the treatment exposure X has an effect on the final outcome Y [2, 1]. We have proposed a finite-population significance Test of the sharp null hypothesis that X has no effect on Y, within the principal stratum of compliers, using a generalized Likelihood Ratio Test [4]. As both the null and alternative hypotheses are composite hypotheses (each comprising a different set of distributions), computing the value of the generalized Likelihood Ratio Test statistic [6] requires two maximizations: one where we assume that the sharp null hypothesis holds, and another without making such an assumption. In our work [4], we have assumed that there are no Always Takers, such that the nuisance parameter is a bivariate parameter describing the total number of Never Takers with observed outcomes y = 0 and y = 1. Extending the approach to the more general case in which there are also Always Takers would require a nuisance parameter of higher dimension that describes the total number of Always Takers with observed outcomes y = 0 and y = 1 as well. This increases the size of the nuisance parameter space and the computational effort needed to find the Likelihood Ratio Test statistic. We present a new algorithm that extends [5] to solve the corresponding integer programs in the general case where there are Always Takers. The procedure for the finite-population significance Test may be illustrated using a toy example from [3].

  • a finite population Likelihood Ratio Test of the sharp null hypothesis for compliers
    Uncertainty in Artificial Intelligence, 2015
    Co-Authors: Wen Wei Loh, Thomas S Richardson
    Abstract:

    In a randomized experiment with noncompliance, scientific interest is often in Testing whether the treatment exposure X has an effect on the final outcome Y. We propose a finite-population significance Test of the sharp null hypothesis that X has no effect on Y, within the principal stratum of Compliers, using a generalized Likelihood Ratio Test. We present a new algorithm that solves the corresponding integer programs.

Jean Jacques Fuchs - One of the best experts on this subject based on the ideXlab platform.

  • the generalized Likelihood Ratio Test and the sparse representations approach
    International Conference on Image and Signal Processing, 2010
    Co-Authors: Jean Jacques Fuchs
    Abstract:

    When sparse representation techniques are used to tentatively recover the true sparse underlying model hidden in an observation vector, they can be seen as solving a joint detection and estimation problem. We consider the l2 - 11 regularized criterion, that is probably the most used in the sparse representation community, and show that, from a detection point of view, minimizing this criterion is similar to applying the Generalized Likelihood Ratio Test. More specifically tuning the regularization parameter in the criterion amounts to set the threshold in the Generalized Likelihood Ratio Test.

  • ICISP - The generalized Likelihood Ratio Test and the sparse representations approach
    Lecture Notes in Computer Science, 2010
    Co-Authors: Jean Jacques Fuchs
    Abstract:

    When sparse representation techniques are used to tentatively recover the true sparse underlying model hidden in an observation vector, they can be seen as solving a joint detection and estimation problem. We consider the l2 - 11 regularized criterion, that is probably the most used in the sparse representation community, and show that, from a detection point of view, minimizing this criterion is similar to applying the Generalized Likelihood Ratio Test. More specifically tuning the regularization parameter in the criterion amounts to set the threshold in the Generalized Likelihood Ratio Test.

Wen Wei Loh - One of the best experts on this subject based on the ideXlab platform.

  • an algorithm to compute the Likelihood Ratio Test statistic of the sharp null hypothesis for compliers
    ACI'15 Proceedings of the UAI 2015 Conference on Advances in Causal Inference - Volume 1504, 2015
    Co-Authors: Wen Wei Loh, Thomas S Richardson
    Abstract:

    In a randomized experiment with noncompliance, scientific interest is often in Testing whether the treatment exposure X has an effect on the final outcome Y [2, 1]. We have proposed a finite-population significance Test of the sharp null hypothesis that X has no effect on Y, within the principal stratum of compliers, using a generalized Likelihood Ratio Test [4]. As both the null and alternative hypotheses are composite hypotheses (each comprising a different set of distributions), computing the value of the generalized Likelihood Ratio Test statistic [6] requires two maximizations: one where we assume that the sharp null hypothesis holds, and another without making such an assumption. In our work [4], we have assumed that there are no Always Takers, such that the nuisance parameter is a bivariate parameter describing the total number of Never Takers with observed outcomes y = 0 and y = 1. Extending the approach to the more general case in which there are also Always Takers would require a nuisance parameter of higher dimension that describes the total number of Always Takers with observed outcomes y = 0 and y = 1 as well. This increases the size of the nuisance parameter space and the computational effort needed to find the Likelihood Ratio Test statistic. We present a new algorithm that extends [5] to solve the corresponding integer programs in the general case where there are Always Takers. The procedure for the finite-population significance Test may be illustrated using a toy example from [3].

  • a finite population Likelihood Ratio Test of the sharp null hypothesis for compliers
    Uncertainty in Artificial Intelligence, 2015
    Co-Authors: Wen Wei Loh, Thomas S Richardson
    Abstract:

    In a randomized experiment with noncompliance, scientific interest is often in Testing whether the treatment exposure X has an effect on the final outcome Y. We propose a finite-population significance Test of the sharp null hypothesis that X has no effect on Y, within the principal stratum of Compliers, using a generalized Likelihood Ratio Test. We present a new algorithm that solves the corresponding integer programs.

Alfred O Hero - One of the best experts on this subject based on the ideXlab platform.

  • measure transformed quasi Likelihood Ratio Test
    International Conference on Acoustics Speech and Signal Processing, 2016
    Co-Authors: Koby Todros, Alfred O Hero
    Abstract:

    In this paper, a generalization of the Gaussian quasi Likelihood Ratio Test (GQLRT) for simple hypotheses is developed. The proposed generalization, called measure-transformed GQLRT (MT-GQLRT), selects a Gaussian probability model that best empirically fits a transformed probability measure of the data. By judicious choice of the transform we show that, unlike the GQLRT, the proposed Test can gain sensitivity to higher-order statistical moments and resilience to outliers leading to significant mitigation of the model mismatch effect on the decision performance. Under some mild regularity conditions we show that the proposed Test statistic is asymptotically normal. A data driven procedure for optimal selection of the measure transformation parameters is developed that maximizes an empirical estimate of the asymptotic power given a fixed empirical asymptotic size. The MT-GQLRT is applied to signal classification in a simulation example that illustrates its sensitivity to higher-order statistical moments and resilience to outliers.

Yoshiro Yamamoto - One of the best experts on this subject based on the ideXlab platform.