Efficient generation of random variates via the ratio-of-uniforms method

Improvements to the conventional ratio-of-uniforms method for random variate generation are proposed. A generalized radio-of-uniforms method is introduced, and it is demonstrated that relocation of the required density via the mode can greatly improve the computational efficiency of the method. We describe a multivariate version of the basic method and summarize a general strategy for efficient ratio-of-uniforms generation. Illustrative examples are given.     

On bivariate transformation of scale distributions

ABSTRACT

Elsewhere, I have promoted (univariate continuous) “transformation of scale” (ToS) distributions having densities of the form 2g(Π^?1(x)) where g is a symmetric distribution and Π is a transformation function with a special property. Here, I develop bivariate (readily multivariate) ToS distributions. Univariate ToS distributions have a transformation of random variable relationship with Azzalini-type skew-symmetric distributions; the bivariate ToS distribution here arises from marginal variable transformation of a particular form of bivariate skew-symmetric distribution. Examples are given, as are basic properties—unimodality, a covariance property, random variate generation—and connections with a bivariate inverse Gaussian distribution are pointed out.     

The expected bit complexity of the von Neumann rejection algorithm

In 1952, von Neumann introduced the rejection method for random variate generation. We revisit this algorithm when we have a source of perfect bits at our disposal. In this random bit model, there are universal lower bounds for generating a random variate with a given density to within an accuracy \(\epsilon \) derived by Knuth and Yao, and refined by the authors. In general, von Neumann’s method fails in this model. We propose a modification that insures proper behavior for all Riemann-integrable densities on compact sets, and show that the expected number of random bits needed behaves optimally with respect to universal lower bounds. In particular, we introduce the notion of an oracle that evaluates the supremum and infimum of a function on any rectangle of \({\mathbb {R}}^{d}\), and develop a quadtree-style extension of the classical rejection method.     

Nonparametric Random Variate Generation Using a Piecewise-Linear Cumulative Distribution Function

The standard approach to solving the interpolation problem for a trace-driven simulation involving a continuous random variable is to construct a piecewise-linear cdf that fills in the gaps between the data values. Some probabilistic properties of this estimator are derived, and three extensions to the standard approach (matching moments, weighted values, and right-censored data) are presented, along with associated random variate generation algorithms. The algorithm is a nonparametric blackbox variate generator requiring only observed data from the user.     

Efficient generation of random vectors by using the ratio-of-uniforms method with ellipsoidal envelopes

Stochastic simulation is widely used to validate procedures and provide guidance for both theoretical and practical problems. Random variate generation is the basis of stochastic simulation. Applying the ratio-of-uniforms method to generate random vectors requires the ability to generate points uniformly in a suitable region of the space. Starting from the observation that, for many multivariate distributions, the multidimensional objective region can be covered by a hyper-ellipsoid more tightly than by a hyper-rectangle, a new algorithm to generate from multivariate distributions is proposed. Due to the computational saving it can produce, this method becomes an appealing statistical tool to generate random vectors from families of standard and nonstandard multivariate distributions. It is particularly interesting to generate from densities known up to a multiplicative constant, for example, from those arising in Bayesian computation. The proposed method is applied and its efficiency is shown for some classes of distributions.     

Random variate generation from D-distributions

Within the context of non-parametric Bayesian inference, Dykstra and Laud (1981) define an extended gamma (EG) process and use it as a prior on increasing hazard rates. The attractive features of the extended gamma (EG) process, among them its capability to index distribution functions that are absolutely continuous, are offset by the intractable nature of the computation that needs to be performed. Sampling based approaches such as the Gibbs Sampler can alleviate these difficulties but the EG processes then give rise to the problem of efficient random variate generation from a class of distributions called D-distributions. In this paper, we describe a novel technique for sampling from such distributions, thereby providing an efficient computation procedure for non-parametric Bayesian inference with a rich class of priors for hazard rates.     

On accurate and precise generation of generalized Poisson variates

This article is concerned with comparison of a few random variate generation techniques for the generalized Poisson distribution. An evaluation is conducted on the degree of proximity between the estimates for its two distributional parameters and first four moments, and the specified or computed true population values via commonly accepted accuracy and precision measures in a simulated setting.     

Sampling from a Discrete Distribution While Preserving Monotonicity

This article describes a cutpoint sampling method for efficiently sampling from an n-point discrete distribution that preserves the monotone relationship between a uniform deviate and the random variate it generates. This property is useful for developing a sampling plan to reduce variance in a Monte Carlo or simulation study. The expected number of comparisons with this method is derived and shown to be bounded above by (m + n ?1)/n, where m denotes the number of cut-points. The alias sampling method, which is regarded as the most efficient table sampling technique, generally lacks the monotone property and requires 2n storage locations, whereas the proposed cutpoint sampling method requires m + n storage locations. The article describes two modifications for cases in which n is large and possibly infinite. It is shown that circumstances arise in which the cutpoint method requires fewer comparisons on average than the alias method does for exactly the same space requirement. The article also describes an algorithm to implement the proposed method.     

On certain unbiased product estimators

When data on an auxiliary variate is available on all the units of the population, negatively correlated with the study variate, Robson (1957) and Murthy (1964) proposed product method of estimation for the estimation of the population total (mean) of the study variate. In this paper, we discuss a method given in Rao (1983) and obtain a simpler dervation of the class of unbiased product estimators for the case of Simple Random Sampling WithOut Replacement design as well as for the case of interpenetrating subsamples design which follows as a limiting case. Finally, we shall illustrate the results by means of two simple numerical example from live data.     

A simple approach for generating correlated binary variates ∗

Correlated binary data arise frequently in medical as well as other scientific disciplines; and statistical methods, such as generalized estimating equation (GEE), have been widely used for their analysis. The need for simulating correlated binary variates arises for evaluating small sample properties of the GEE estimators when modeling such data. Also, one might generate such data to simulate and study biological phenomena such as tooth decay or periodontal disease. This article introduces a simple method for generating pairs of correlated binary data. A simple algorithm is also provided for generating an arbitrary dimensional random vector of non-negatively correlated binary variates. The method relies on the idea that correlations among the random variables arise as a result of their sharing some common components that induce such correlations. It then uses some properties of the binary variates to represent each variate in terms of these common components in addition to its own elements. Unlike most previous approaches that require solving nonlinear equations or use some distributional properties of other random variables, this method uses only some properties of the binary variate. As no intermediate random variables are required for generating the binary variates, the proposed method is shown to be faster than the other methods. To verify this claim, we compare the computational efficiency of the proposed method with those of other procedures.     

Extending Product Method of Estimation to Positive Correlation Case in Surveys

The product method of estimation (Murthy, 1964) complements the ratio method when the study variate, y, and an auxiliary variate, x, have negative correlation. However, such cases are not frequent in survey practice. This paper suggests a simple transformation of x in the more common situation of positive correlation between y and x, to permit a product method of estimation rather than a ratio method. This leads to the advantage that the bias and mean square error have exact expressions. The technique developed by Quenouille (1956) and applied by Shukla (1976) is used for making the estimator unbiased. The minimum variance situation is investigated. Two numerical examples are included. The case of negative correlation is also examined.     

On random walks on affine group

Diaconis' presumption that the number of steps required to get close to uniform for a random walk on the affine group A _pis c(p)p ²with c(p) →ã is verified. We also discuss the random number generation associated with the random walk on the affine group. The number of steps to force the generated number to become random is improved. A modified version of Diacohis-Shahshahani's upper bound lemma is given and applied     

Estimating the number of clusters

Hartigan (1975) defines the number q of clusters in a d ‐variate statistical population as the number of connected components of the set {f > c}, where f denotes the underlying density function on R^d and c is a given constant. Some usual cluster algorithms treat q as an input which must be given in advance. The authors propose a method for estimating this parameter which is based on the computation of the number of connected components of an estimate of {f > c}. This set estimator is constructed as a union of balls with centres at an appropriate subsample which is selected via a nonparametric density estimator of f. The asymptotic behaviour of the proposed method is analyzed. A simulation study and an example with real data are also included.     

Erratum

For a random variable obeying the inverse Gaussian distribu-tion and its reciprocal, the uniformly minimum variance unbiased (UMVU) estimators of each mode are obtained. The UMVU estimators

of the left and right limits of a certain interval which contains an inverse Gaussian variate with an arbitrary given probability are also proposed.     

A simple generalization of the Box–Muller method for obtaining a pair of correlated standard normal variables

In the simple and widely used method of Box–Muller [G. Box and M. Muller, A note on the generation of random normal deviates, Ann. Math. Statist. 29 (1958), pp. 610–611], from a pair of uniform and independent random variables in (0,1), a pair of standard and independent normal variables is obtained. In this article, we present a very simple and elegant generalization of this method to obtain a pair of correlated standard normal variables with a given coefficient of correlation. This generalized method, which is computationally very easy, is interpreted in geometric terms, considering a translation of the uniform interval (0,1) and a rotation of a defined angle, both related to the coefficient of correlation. Some numerical results are simulated and statistically analysed, proving that the generalization is extremely simple and powerful.     

Layer Sampling

Layer sampling is an algorithm for generating variates from a non-normalized multidimensional distribution p( · ). It empirically constructs a majorizing function for p( · ) from a sequence of layers. The method first selects a layer based on the previous variate. Next, a sample is drawn from the selected layer, using a method such as Rejection Sampling. Layer sampling is regenerative. At regeneration times, the layers may be adapted to increase mixing of the Markov chain. Layer sampling may also be used to estimate arbitrary integrals, including normalizing constants.     

The inverse problem of multivariate and matrix-variate skew normal distributions

In this paper, we prove that the joint distribution of random vectors Z ₁ and Z ₂ and the distribution of Z ₂ are skew normal provided that Z ₁ is skew normally distributed and Z ₂ conditioning on Z ₁ is distributed as closed skew normal. Also, we extend the main results to the matrix variate case.     

Critical values of some nonparametric statistics used for tests of total stochastic independence

The problem of testing for total independence of the variates of a stochastic p(≧3) component vector using rank correlation statistics is considered. Two distribution free statistics are considered, one based on the determinant of the matrix of rank correlation statistics, the second on their sum of squares. Tables of critical values are given for p=3,4 for the cases when (a) ranks, and (b) exponential scores are used to replace the ordered observations within each variate. Some approximations to the critical values are proposed and evaluated.     

Generalized Linear Models

The importance of random number generators has increased over the years. This follows from the fact that contemporary research methods rely more and more on simulation and the increased importance of encryption technology. The output of a random number generator is created by either an algorithm or a physical device. The most popular method for random number generation is through the use of an algorithm. This article presents a new category of physical random bit generator that is packaged by several manufacturers. A statistical analysis of the output from the generators is given.     

A Sample Size Formula for a Non-Central t Test

An elementary method of proof of the mode, median, and mean inequality is given for skewed, unimodal distributions of continuous random variables. A proof of the inequality for the gamma, F, and beta random variables is sketched.