-
Random Discrete Probability Measures Based on Negative Binomial Process
Authors:
Sadegh Chegini,
Mahmoud Zarepour
Abstract:
An important functional of Poisson random measure is the negative binomial process (NBP). We use NBP to introduce a generalized Poisson-Kingman distribution and its corresponding random discrete probability measure. This random discrete probability measure provides a new set of priors with more flexibility in nonparametric Bayesian models. It is shown how this random discrete probability measure r…
▽ More
An important functional of Poisson random measure is the negative binomial process (NBP). We use NBP to introduce a generalized Poisson-Kingman distribution and its corresponding random discrete probability measure. This random discrete probability measure provides a new set of priors with more flexibility in nonparametric Bayesian models. It is shown how this random discrete probability measure relates to the non-parametric Bayesian priors such as Dirichlet process, normalized positive α-stable process, Poisson-Dirichlet process (PDP), and others. An extension of the DP with its almost sure approximation is presented. Using our representation for NBP, we derive a new series representation for the PDP.
△ Less
Submitted 30 June, 2023;
originally announced July 2023.
-
A Note on Bayesian Nonparametric Inference for Spherically Symmetric Distribution
Authors:
Reyhaneh Hosseini,
Mahmoud Zarepour
Abstract:
In this paper, we describe a Bayesian nonparametric approach to make inference for a bivariate spherically symmetric distribution. We consider a Dirichlet invariant process prior on the set of all bivariate spherically symmetric distributions and we derive the Dirichlet invariant process posterior. Indeed, our approach is an extension of Dirichlet invariant process for the symmetric distributions…
▽ More
In this paper, we describe a Bayesian nonparametric approach to make inference for a bivariate spherically symmetric distribution. We consider a Dirichlet invariant process prior on the set of all bivariate spherically symmetric distributions and we derive the Dirichlet invariant process posterior. Indeed, our approach is an extension of Dirichlet invariant process for the symmetric distributions on the real line to a bivariate spherically symmetric distribution where the underlying distribution is invariant under a finite group of rotations. Moreover, we obtain the Dirichlet invariant process posterior for the infinite transformation group and we prove that it approaches to Dirichlet process.
△ Less
Submitted 31 July, 2018; v1 submitted 29 July, 2018;
originally announced July 2018.
-
A Note on Bootstrapping M-estimates from Unstable AR(2) Process with Infinite Variance Innovations
Authors:
Maryam Sohrabi,
Mahmoud Zarepour
Abstract:
The limiting distribution for M-estimates in a non-stationary autoregressive model with heavy-tailed error is computationally intractable. To make inferences based on the M-estimates, the bootstrap procedure can be used to approximate the sampling distribution. In this paper, we show that the bootstrap scheme with $m=o(n)$ resampling sample size when $m/n \to 0$ is approximately valid in a multipl…
▽ More
The limiting distribution for M-estimates in a non-stationary autoregressive model with heavy-tailed error is computationally intractable. To make inferences based on the M-estimates, the bootstrap procedure can be used to approximate the sampling distribution. In this paper, we show that the bootstrap scheme with $m=o(n)$ resampling sample size when $m/n \to 0$ is approximately valid in a multiple unit roots time series with innovations in the domain of attraction of a stable law with index $0<α\leq2$.
△ Less
Submitted 8 March, 2016;
originally announced March 2016.
-
A Bayesian nonparametric chi-squared goodness-of-fit test
Authors:
Reyhaneh Hosseini,
Mahmoud Zarepour
Abstract:
The Bayesian nonparametric inference and Dirichlet process are popular tools in statistical methodologies. In this paper, we employ the Dirichlet process in hypothesis testing to propose a Bayesian nonparametric chi-squared goodness-of-fit test. In our Bayesian nonparametric approach, we consider the Dirichlet process as the prior for the distribution of data and carry out the test based on the Ku…
▽ More
The Bayesian nonparametric inference and Dirichlet process are popular tools in statistical methodologies. In this paper, we employ the Dirichlet process in hypothesis testing to propose a Bayesian nonparametric chi-squared goodness-of-fit test. In our Bayesian nonparametric approach, we consider the Dirichlet process as the prior for the distribution of data and carry out the test based on the Kullback-Leibler distance between the updated Dirichlet process and the hypothesized distribution F0. We prove that this distance asymptotically converges to the same chi-squared distribution as the chi-squared test does. Similarly, a Bayesian nonparametric chi-squared test of independence for a contingency table is provided. Also, by computing the Kullback-Leibler distance between the Dirichlet process and the hypothesized distribution, a method to obtain an appropriate concentration parameter for the Dirichlet process is suggested.
△ Less
Submitted 16 June, 2016; v1 submitted 30 January, 2016;
originally announced February 2016.
-
Asymptotic Theory for M-Estimates in Unstable AR(p) Processes with Infinite Variance Innovations
Authors:
Maryam Sohrabi,
Mahmoud Zarepour
Abstract:
In this paper, we present the asymptotic distribution of M-estimators for parameters in non-stationary AR(p) processes. The innovations are assumed to be in the domain of attraction of a stable law with index $0<α\le2$. In particular, when the model involves repeated unit roots or conjugate complex unit roots, M-estimators have a higher asymptotic rate of convergence compared to the least square e…
▽ More
In this paper, we present the asymptotic distribution of M-estimators for parameters in non-stationary AR(p) processes. The innovations are assumed to be in the domain of attraction of a stable law with index $0<α\le2$. In particular, when the model involves repeated unit roots or conjugate complex unit roots, M-estimators have a higher asymptotic rate of convergence compared to the least square estimators and the asymptotic results can be written as Itô stochastic integrals.
△ Less
Submitted 12 December, 2016; v1 submitted 18 June, 2015;
originally announced June 2015.
-
On Approximations of the Beta Process in Latent Feature Models
Authors:
Luai Al Labadi,
Mahmoud Zarepour
Abstract:
The beta process has recently been widely used as a nonparametric prior for different models in machine learning, including latent feature models. In this paper, we prove the asymptotic consistency of the finite dimensional approximation of the beta process due to Paisley \& Carin (2009). In addition, we derive an almost sure approximation of the beta process. This approximation provides a direct…
▽ More
The beta process has recently been widely used as a nonparametric prior for different models in machine learning, including latent feature models. In this paper, we prove the asymptotic consistency of the finite dimensional approximation of the beta process due to Paisley \& Carin (2009). In addition, we derive an almost sure approximation of the beta process. This approximation provides a direct method to efficiently simulate the beta process. A simulated example, illustrating the work of the method and comparing its performance to several existing algorithms, is also included.
△ Less
Submitted 12 November, 2014;
originally announced November 2014.
-
Two-sample Bayesian nonparametric goodness-of-fit test
Authors:
Luai Al Labadi,
Emad Masuadi,
Mahmoud Zarepour
Abstract:
In recent years, Bayesian nonparametric statistics has gathered extraordinary attention. Nonetheless, a relatively little amount of work has been expended on Bayesian nonparametric hypothesis testing. In this paper, a novel Bayesian nonparametric approach to the two-sample problem is established. Precisely, given two samples $\mathbf{X}=X_1,\ldots,X_{m_1}$ $\overset {i.i.d.} \sim F$ and…
▽ More
In recent years, Bayesian nonparametric statistics has gathered extraordinary attention. Nonetheless, a relatively little amount of work has been expended on Bayesian nonparametric hypothesis testing. In this paper, a novel Bayesian nonparametric approach to the two-sample problem is established. Precisely, given two samples $\mathbf{X}=X_1,\ldots,X_{m_1}$ $\overset {i.i.d.} \sim F$ and $\mathbf{Y}=Y_1,\ldots,Y_{m_2} \overset {i.i.d.} \sim G$, with $F$ and $G$ being unknown continuous cumulative distribution functions, we wish to test the null hypothesis $\mathcal{H}_0:~F=G$. The method is based on the Kolmogorov distance and approximate samples from the Dirichlet process centered at the standard normal distribution and a concentration parameter 1. It is demonstrated that the proposed test is robust with respect to any prior specification of the Dirichlet process. A power comparison with several well-known tests is incorporated. In particular, the proposed test dominates the standard Kolmogorov-Smirnov test in all the cases examined in the paper.
△ Less
Submitted 7 May, 2015; v1 submitted 12 November, 2014;
originally announced November 2014.
-
On Some Asymptotic Properties and an Almost Sure Approximation of the Normalized Inverse-Gaussian Process
Authors:
Luai Al Labadi,
Mahmoud Zarepour
Abstract:
In this paper, we present some asymptotic properties of the normalized inverse-Gaussian process. In particular, when the concentration parameter is large, we establish an analogue of the empirical functional central limit theorem, the strong law of large numbers and the Glivenko-Cantelli theorem for the normalized inverse-Gaussian process and its corresponding quantile process. We also derive a fi…
▽ More
In this paper, we present some asymptotic properties of the normalized inverse-Gaussian process. In particular, when the concentration parameter is large, we establish an analogue of the empirical functional central limit theorem, the strong law of large numbers and the Glivenko-Cantelli theorem for the normalized inverse-Gaussian process and its corresponding quantile process. We also derive a finite sum-representation that converges almost surely to the Ferguson and Klass representation of the normalized inverse-Gaussian process. This almost sure approximation can be used to simulate efficiently the normalized inverse-Gaussian process.
△ Less
Submitted 28 June, 2012;
originally announced June 2012.
-
The Dirichlet Process with Large Concentration Parameter
Authors:
Luai Al Labadi,
Mahmoud Zarepour
Abstract:
Ferguson's Dirichlet process plays an important role in nonparametric Bayesian inference. Let $P_a$ be the Dirichlet process in $\mathbb{R}$ with a base probability measure $H$ and a concentration parameter $a>0.$ In this paper, we show that $\sqrt {a} \big(P_a((-\infty,t]) -H((-\infty,t])\big)$ converges to a certain Brownian bridge as $a \to \infty.$ We also derive a certain Glivenko-Cantelli th…
▽ More
Ferguson's Dirichlet process plays an important role in nonparametric Bayesian inference. Let $P_a$ be the Dirichlet process in $\mathbb{R}$ with a base probability measure $H$ and a concentration parameter $a>0.$ In this paper, we show that $\sqrt {a} \big(P_a((-\infty,t]) -H((-\infty,t])\big)$ converges to a certain Brownian bridge as $a \to \infty.$ We also derive a certain Glivenko-Cantelli theorem for the Dirichlet process. Using the functional delta method, the weak convergence of the quantile process is also obtained. A large concentration parameter occurs when a statistician puts too much emphasize on his/her prior guess. This scenario also happens when the sample size is large and the posterior is used to make inference.
△ Less
Submitted 13 December, 2011; v1 submitted 24 September, 2011;
originally announced September 2011.
-
Random probability measures via Polya sequences: revisiting the Blackwell-MacQueen urn scheme
Authors:
Hemant Ishwaran,
Mahmoud Zarepour
Abstract:
Sufficient conditions are developed for a class of generalized Polya urn schemes ensuring exchangeability. The extended class includes the Blackwell-MacQueen Polya urn and the urn schemes for the two-parameter Poisson-Dirichlet process and finite dimensional Dirichlet priors among others.
Sufficient conditions are developed for a class of generalized Polya urn schemes ensuring exchangeability. The extended class includes the Blackwell-MacQueen Polya urn and the urn schemes for the two-parameter Poisson-Dirichlet process and finite dimensional Dirichlet priors among others.
△ Less
Submitted 2 September, 2003;
originally announced September 2003.
