Results 1  10
of
12
Toward Simplifying and Accurately Formulating Fragment Assembly
 JOURNAL OF COMPUTATIONAL BIOLOGY
, 1995
"... The fragment assembly problem is that of reconstructing a DNA sequence from a collection of randomly sampled fragments. Traditionally the objective of this problem has been to produce the shortest string that contains all the fragments as substrings, but in the case of repetitive target sequence ..."
Abstract

Cited by 37 (1 self)
 Add to MetaCart
The fragment assembly problem is that of reconstructing a DNA sequence from a collection of randomly sampled fragments. Traditionally the objective of this problem has been to produce the shortest string that contains all the fragments as substrings, but in the case of repetitive target sequences this objective produces answers that are overcompressed. In this paper, the problem is reformulated as one of finding a maximumlikelihood reconstruction with respect to the 2sided KolmogorovSmirnov statistic, and it is argued that this is a better formulation of the problem. Next the fragment assembly problem is recast in graphtheoretic terms as one of finding a noncyclic subgraph with certain properties and the objectives of being shortest or maximallylikely are also recast in this framework. Finally, a series of graph reduction transformations are given that dramatically reduce the size of the graph to be explored in practical instances of the problem. This reduction is ...
Statistical Models for Automatic Performance Tuning
 In Proceedings of the 2001 International Conference on Computational Science (ICCS 2001
, 2001
"... Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have emerged in response, and they typically operate by (1) generating a large number of possible implementations of a subroutine, and (2) selecting the fast ..."
Abstract

Cited by 16 (4 self)
 Add to MetaCart
Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have emerged in response, and they typically operate by (1) generating a large number of possible implementations of a subroutine, and (2) selecting the fastest implementation by an exhaustive, empirical search. This paper presents quantitative data that motivates the development of such a searchbased system, and discusses two problems which arise in the context of search. First, we develop a heuristic for stopping an exhaustive compiletime search early if a nearoptimal implementation is found. Second, we show how to construct runtime decision rules, based on runtime inputs, for selecting from among a subset of the best implementations.
Statistical modeling of signal amplitude fading of indoor radiopropagation channels
 In Proceedings of the 2nd International Conference on Universal Personal Communications 1
, 1993
"... Smallscale and largescale variations of signal amplitude for wideband data collected at two office buildings have been investigated. The data base includes 12000 impulse response estimates which were obtained by inverse Fourier transforming of the channel's frequency response profiles. Extensive c ..."
Abstract

Cited by 6 (0 self)
 Add to MetaCart
Smallscale and largescale variations of signal amplitude for wideband data collected at two office buildings have been investigated. The data base includes 12000 impulse response estimates which were obtained by inverse Fourier transforming of the channel's frequency response profiles. Extensive curve fitting for the distribution of individual multipath components' amplitudes using elaborate tests based on the KolmogorovSmirnov and Wilcoxon procedures shows that amplitude fading is lognormal for both local and global data. For the local data, the lognormal, Nakagami, and Rayleigh distributions passed the KolmogorovSmirnov test with a 90 % confidence level for 76%, 3696, and 13 % of cases, respectively. For global data the Wilcoxon test showed that with 99 % confidence lognormal is better than Nakagami, and with 94 % confidence Nakagami fits the data better than Rayleigh. After establishing the lognormality of signal amplitude fading in these environments, a twodimensional Gaussian distrihution for simulating the log amplitudes of spatiallyadjacent profiles has been proposed and implemented.
The modelsize effect on traditional and modified tests of covariance structures
 Structural Equation Modeling
, 2007
"... According to Kenny and McCoach (2003), chisquare tests of structural equation models produce inflated Type I error rates when the degrees of freedom increase. So far, the amount of this bias in large models has not been quantified. In a Monte Carlo study of confirmatory factor models with a range o ..."
Abstract

Cited by 5 (4 self)
 Add to MetaCart
According to Kenny and McCoach (2003), chisquare tests of structural equation models produce inflated Type I error rates when the degrees of freedom increase. So far, the amount of this bias in large models has not been quantified. In a Monte Carlo study of confirmatory factor models with a range of 48 to 960 degrees of freedom it was found that the traditional maximum likelihood ratio statistic, TML, overestimates nominal Type I error rates up to 70 % under conditions of multivariate normality. Some alternative statistics for the correction of modelsize effects were also investigated: the scaled Satorra–Bentler statistic, TSC; the adjusted Satorra–
Statistical Modeling of Feedback Data in an Automatic Tuning System
 in MICRO33: Third ACM Workshop on FeedbackDirected Dynamic Optimization
, 2000
"... Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have been developed in response which typically operate, at compiletime, by (1) generating a large number of possible implementations of a subroutine, and (2) s ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have been developed in response which typically operate, at compiletime, by (1) generating a large number of possible implementations of a subroutine, and (2) selecting a fast implementation by an exhaustive, empirical search. In this paper, we show how statistical modeling of the performance feedback data collected during the search phase can be used in two novel and important ways. First, we develop a heuristic for stopping an exhaustive compiletime search early if a nearoptimal implementation is found. Second, we show how to construct runtime decision rules, based on runtime inputs, for selecting from among a subset of the best implementations. We apply our methods to actual performance data collected by the PHiPAC tuning system for matrix multiply on a variety of hardware and compiler platforms.
Computing the Joint Distribution of General Linear Combinations of Spacings or Exponential Variates
 Sinica
, 2001
"... We present an algorithm for computing exact expressions for the distribution of the maximum or minimum of an arbitrary finite collection of linear combinations of spacings or exponential random variables with rational coefficients. These expressions can then be manipulated or evaluated using symb ..."
Abstract

Cited by 3 (1 self)
 Add to MetaCart
We present an algorithm for computing exact expressions for the distribution of the maximum or minimum of an arbitrary finite collection of linear combinations of spacings or exponential random variables with rational coefficients. These expressions can then be manipulated or evaluated using symbolic math packages such as MAPLE. As examples, we apply this algorithm to obtain the distributions of the maximum and minimum of a moving average process, and the distribution of the KolmogorovSmirnov statistic. Key words and phrases: KolmogorovSmirnov statistic, moving average process, symbolic computations. 1 Introduction Let S (n) denote the vector of spacings between n random points on the interval (0; 1). More precisely, suppose that X 1 ; X 2 ; : : : ; X n are i.i.d. from a uniform distribution on the interval (0, 1), and let X (1) X (2) \Delta \Delta \Delta X (n) be the corresponding order statistics. We define the spacings S 1 ; S 2 ; : : : ; S n+1 to be the successive di...
Some New Test Statistics for Mean and Covariance Structure Analysis with High Dimensional Data
"... Covariance structure analysis is often used for inference and for dimension reduction with high dimensional data. When data is not normally distributed, the asymptotic distribution free (ADF) method is often used to fit a proposed model. This approach uses a weight matrix based on the inverse of the ..."
Abstract
 Add to MetaCart
Covariance structure analysis is often used for inference and for dimension reduction with high dimensional data. When data is not normally distributed, the asymptotic distribution free (ADF) method is often used to fit a proposed model. This approach uses a weight matrix based on the inverse of the matrix formed by the sample fourth moments and sample covariances. The ADF test statistic is asymptotically distributed as a chisquare variate, but its empirical performance rejects the true model too often at all but impractically large sample sizes. By comparing mean and covariance structure analysis with its peer in the multivariate linear model, we propose some modified ADF test statistics as Ftests whose distributions we approximate using Fdistributions. Empirical studies show that the distributions of the new Ftests are more closely approximated by Fdistributions than are the original ADF statistics when referred to chisquare distributions. Detailed analysis indicates why the AD...
On Statistical Models in Automatic Tuning
"... . Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have emerged in response, and they typically operate, at compiletime, by (1) generating a large number of possible implementations of a subroutine, and (2 ..."
Abstract
 Add to MetaCart
. Achieving peak performance from library subroutines usually requires extensive, machinedependent tuning by hand. Automatic tuning systems have emerged in response, and they typically operate, at compiletime, by (1) generating a large number of possible implementations of a subroutine, and (2) selecting a fast implementation by an exhaustive, empirical search. This paper applies statistical techniques to exploit the large amount of performance data collected during the search. First, we develop a heuristic for stopping an exhaustive compiletime search early if a nearoptimal implementation is found. Second, we show how to construct runtime decision rules, based on runtime inputs, for selecting from among a subset of the best implementations. We apply our methods to actual performance data collected by the PHiPAC tuning system for matrix multiply on a variety of hardware platforms. 1
Exchange Intervention
"... More than 20 percent of the funds that banks have on deposit with the Federal Reserve Banks are required clearing balances, not required reserve balances. Since 1981, when they first earned a market return, clearing balances have become widespread among banks of all sizes. Here, the author takes a l ..."
Abstract
 Add to MetaCart
More than 20 percent of the funds that banks have on deposit with the Federal Reserve Banks are required clearing balances, not required reserve balances. Since 1981, when they first earned a market return, clearing balances have become widespread among banks of all sizes. Here, the author takes a look at the reasons for the popularity of this relatively new phenomenon as well as its impact on the setting and measuring of monetary policy.