Heavy - Tailed Probability Distributions
نویسندگان
چکیده
Combinatorial search methods often exhibit a large variability in performance. We study the cost prooles of combinatorial search procedures. Our study reveals some intriguing properties of such cost prooles. The distributions are often characterized by very long tails or \heavy tails". We will show that these distributions are best characterized by a general class of distributions that have no moments (i.e., an innnite mean, variance , etc.). Such non-standard distributions have recently been observed in areas as diverse as economics, statistical physics, and geophysics. They are closely related to fractal phenomena, whose study was introduced by Mandelbrot. We believe this is the rst nding of these distributions in a purely computational setting. We also show how random restarts can effectively eliminate heavy-tailed behavior, thereby dramatically improving the overall performance of a search procedure. This paper has not already been accepted by and is not currently under review for another conference. Nor will it be submitted for such during CP's review period.
منابع مشابه
Open Problem: Restricted Eigenvalue Condition for Heavy Tailed Designs
The restricted eigenvalue (RE) condition characterizes the sample complexity of accurate recovery in the context of high-dimensional estimators such as Lasso and Dantzig selector (Bickel et al., 2009). Recent work has shown that random design matrices drawn from any thin-tailed (subGaussian) distributions satisfy the RE condition with high probability, when the number of samples scale as the sq...
متن کاملHow to Define Mean, Variance, etc., for Heavy-Tailed Distributions: A Fractal-Motivated Approach
In many practical situations, we encounter heavy-tailed distributions for which the variance – and even sometimes the mean – are infinite. We propose a fractal-based approach that enables us to gauge the mean and variance of such distributions. 1 Formulation of the Problem There are many practical situations in which the probability distribution is drastically different from normal. In many suc...
متن کاملHeavy-tailed random matrices
We discuss non-Gaussian random matrices whose elements are random variables with heavy-tailed probability distributions. In probability theory heavy tails of the distributions describe rare but violent events which usually have dominant influence on the statistics. They also completely change universal properties of eigenvalues and eigenvectors of random matrices. We concentrate here on the uni...
متن کاملMeasures of Deviation (and Dependence) for Heavy-Tailed Distributions and their Estimation under Interval and Fuzzy Uncertainty
Traditionally, in science and engineering, most statistical techniques are based on the assumption that the random variables are normally distributed. For such distributions, a natural characteristic of the “average” value is the mean, and a natural characteristic of the deviation from the average is the variance. However, in many practical situations, e.g., in economics and finance, we encount...
متن کاملNumerical Methods for Analyzing Queues with Heavy-Tailed Distributions
In many Internet-type queues, arrival and service distributions are heavy-tailed. A difficulty with analyzing these queues is that heavy-tailed distributions do not generally have closed-form Laplace transforms. A recently proposed method, the Transform Approximation Method (TAM), overcomes this by numerically approximating the transform. This paper investigates numerical issues of implementing...
متن کاملNeed for Most Accurate Discrete Approximations Explains Heavy-Tailed Distributions
In many practical situations, we encounter Gaussian distributions, for which the distribution tails are light – in the sense that as the value increases, the corresponding probability density tends to 0 very fast. There are many theoretical explanations for the Gaussian distributions and for similar light-tail distributions. In practice, however, we often encounter heavy-tailed distributions, i...
متن کامل