Stochastic Storage Processes: Queues, Insurance Risk, Dams, and Data Communication (Stochastic Modelling and Applied Probability) by N.U. Prabhu
English | Sep. 27, 2012 | ISBN: 1461272602 | 217 Pages | PDF | 6 MB

A self-contained treatment of stochastic processes arising from models for queues, insurance risk, and dams and data communication, using their sample function properties. The approach is based on the fluctuation theory of random walks, L vy processes, and Markov-additive processes, in which Wiener-Hopf factorisation plays a central role.

Jiti Gao, "Nonlinear Time Series: Semiparametric and Nonparametric Methods"
English | ISBN: 1584886137 | 2007 | 237 pages | PDF | 5 MB

Nolberto Solano, now Peru’s assistant manager, knows exactly what’s at stake.

“We can lose against anyone, but less so against Chile because it is a Clásico,” he bawls in this handy preview of tonight’s affair by my colleague Gerard Meagher.

11.31pm BST

• Left-back reaps reward for outstanding first season in top flight
• ‘It’s going to be a massive year for us, the last year at Upton Park’
Aaron Cresswell has signed a new five-year contract at West Ham after playing a starring role in his debut season.

The left-back was signed from Ipswich last summer and went on to play every game in West Ham’s Premier League campaign, his first in the top-flight.

Bombing comes a week after another attack that government said may have caused the worst environmental disaster in the country’s history

Colombia’s Farc guerrillas bombed a new section of an oil pipeline on Monday, even as cleanup teams struggled to contain a massive oil spill that the government said may be the worst environmental disaster in the country’s history.

A bomb ripped through part of the Tansandino pipeline in southern Putumayo province before dawn on Monday, damaging two homes. Minimal oil was spilled because the pipeline was inactive at the time, officials said.

Metropolitan police arrest man after assault in grounds of Blackfriars crown court

Detectives from the Metropolitan police’s homicide and major crime command are investigating the serious assault of a female Serco security guard at a London court.

“At approximately 1.35pm on Monday 29 June, officers were alerted to a Serco security guard who had been seriously assaulted by a male prisoner at Blackfriars crown court,” the Met said in a statement.

Al Jazeera’s digital news channel AJ+ has released a video report profiling LGBT rights activist Yanzi Peng, a gay man (also known as Yang Teng and Xiao Zhen in earlier reports) who last year won a landmark court case against the gay conversion clinic that had administered electric shock therapy on him in attempt to change his sexual orientation:

Counter-Terrorism Act, which also applies to NHS trusts, schools and further education institutions, comes into force

Local authorities, prisons, NHS trusts, schools, universities and further education institutions will this week be placed under a new statutory duty to prevent extremist radicalisation taking place within their walls.

The requirement was imposed by this year’s Counter-Terrorism and Security Act, and Home Office ministers have pointed out that the wide-ranging powers come into force in the week that David Cameron demanded “a full spectrum response” to the killing of as many as 30 British tourists in Tunisia.

View gallery

A law enforcement officer walks in front of an ambulance at the Alice Hyde Medical Center in Malone, …
By Laila Kearney

NEW YORK (Reuters) - Doctors updated the condition of New York prison escapee David Sweat to serious from critical on Monday, a day after he was shot by police after more than three weeks on the run, hospital officials said.

Sweat, 35, is being treated in the state capital where a trauma team at Albany Medical Center determined that the convict did not require surgery, the hospital said in a statement.

The Supreme Court is refusing to allow Texas to enforce restrictions that would force 10 abortion clinics to close.

The justices voted 5-4 on Monday granting an emergency appeal from the clinics after a federal appeals court upheld new regulations and refused to keep them on hold while the clinics appealed to the Supreme Court.

The Supreme Court order will remain in effect at least until the court decides whether to hear the clinics’ appeal of the lower court ruling, not before the fall.

This is the biggest case on the hot-button subject in nearly a quarter-century.

View photo

A general view of the U.S. Supreme Court building in Washington June 15, 2015. REUTERS/Carlos Barria
By Lawrence Hurley

WASHINGTON (Reuters) - The U.S. Supreme Court on Monday ruled that a drug used by Oklahoma as part of its lethal injection procedure does not violate the U.S. Constitution's ban on cruel and unusual punishment, dealing a setback to opponents of the death penalty.

The court, in a 5-4 decision, handed a loss to three inmates who objected to the use of a sedative called midazolam, saying it cannot achieve the level of unconsciousness required for surgery, making it unsuitable for executions.

Justice Samuel Alito wrote on behalf of the court’s conservative majority that the inmates had, among other things, failed to show that there was an alternative method of execution available that would be less painful.

In a dissenting opinion, liberal Justice Stephen Breyer said the court should consider whether the death penalty itself is constitutional. He was joined by one of his colleagues, Justice Ruth Bader Ginsburg.

Forbes magazine’s annual Celeb 100 list, chronicling the world’s highest-paid entertainers over the past year featured a wide range of talents from all over the globe including a British chef, a Canadian rapper and a Russian tennis star – but only 16 women.

Escaped captives told BBC Panorama that some of the 219 girls flog those who can’t recite Qur’an and slit throats of captured males

Some of the Nigerian schoolgirls kidnapped by Boko Haram militants last year have been brainwashed to begin fighting for the Islamist group, with many carrying out public beatings and even killings, other captives have told the BBC.

Women who claim they lived in the same camps as some of the 219 girls who were taken from their school in the town of Chibok last April told the Panorama progamme that many are now administering punishments on behalf of Boko Haram.

Continue reading...

Here’s what Our Man Wilson has gone and got himself into this evening:

Above the memorial, a solid block of red. pic.twitter.com/a7yUynyC01

12.02am BST

Mike MacKenzie is rooting for the underdog, and doesn’t care who knows it:

“It would be great to see a Peru - Paraguay final but that still seems rather fanciful. More likely it will be Argentina vs the hosts. In that case, I hope Messi and friends win to help silence the ‘what has he won with his country’ critics.”

Continue reading...

Although the parameters in a finite mixture model are unidentifiable, there is a form of local identifiability guaranteeing the existence of the identifiable parameter regions. To verify its existence, practitioners use the Fisher information on the estimated parameters. However, there exist model/data situations where local identifiability based on Fisher information does not correspond to that based on the likelihood. In this paper, we propose a method to empirically measure degree of local identifiability on the estimated parameters, empirical identifiability, based on one’s ability to construct an identifiable likelihood set. From a detailed topological study of the likelihood region, we show that for any given data set and mixture model, there typically exists limited range of confidence levels where the likelihood region has a natural partition into identifiable subsets. At confidence levels that are too high, there is no natural way to use the likelihood to resolve the identifiability problem.

We consider estimation of the quadratic (co)variation of a semimartingale from discrete observations which are irregularly spaced under high-frequency asymptotics. In the univariate setting, results by Jacod for regularly spaced observations are generalized to the case of irregular observations. In the two-dimensional setup under non-synchronous observations, we derive a stable central limit theorem for the Hayashi–Yoshida estimator in the presence of jumps. We reveal how idiosyncratic and simultaneous jumps affect the asymptotic distribution. Observation times generated by Poisson processes are explicitly discussed.

Jim Cochran


In the sixth installment of the Amstat News series of interviews with ASA presidents and executive directors, we feature a discussion with 2003 ASA President J. Stuart (Stu) Hunter.

Stuart_HunterJ. Stuart Hunter is a statistician and professor emeritus in the school of engineering and applied science at Princeton University. He earned a bachelor’s degree in electrical engineering in 1947, a master’s degree in applied mathematics in 1949, and a PhD in statistics in 1954—all from North Carolina State University. He is the founding editor of Technometrics and, in 1993, he served as president of the American Statistical Association. He is a fellow of the American Statistical Association, American Society for Quality, Royal Statistical Society, and American Association for the Advancement of Science. Hunter became an honorary member of the American Society for Quality in 1999 and he was elected a member of the National Academy of Engineering in 2005.
The first “Stu Hunter Research Conference” was held in 2013 at the Chateau Marquette, in Amsterdam, the Netherlands. The second of these conferences was held in Phoenix, Arizona, in 2014, and the third was held in March of 2015 in Leuven, Belgium.

Hunter is a highly decorated statistician. He received the Shewhart Medal in 1970, U.S. Army S. S. Wilks Medal in 1987, Deming Medal in 1986, and Founders Award of the American Statistical Association in 1995. He also has been honored with the W.J. Youden, Ellis Ott, and Brunbaugh awards of the American Society for Quality. He received an honorary degree from his alma mater in 2006 and again in 2008 from The Pennsylvania State University. In 1996, the Environmetrics Society established an annual lecture in his honor.

Hunter has published extensively and is a co-author with I. Guttman and S. S. Wilks of Introductory Engineering Statistics (1965) and co-author with
G. E. P. Box and W. G. Hunter of Statistics for Experimenters (1979). In 1968, he was the instructor in the 32 one-half hour episode TV course “The Design of Experimenters with Structured Text” and, in 1971, a similar program, “Statistics for Problem Solving and Decision Making,” all sponsored by Westinghouse Learning. He remains active as a consultant and lecturer.

Q: Stu, thank you for taking time for this interview. You earned your undergraduate degree in electrical engineering and you master’s degree in engineering mathematics, but then studied statistics for your doctorate. What factored into your decision to move into the field of statistics?

A: In 1940, after graduating from high school, I worked for the Prudential Insurance Company as junior clerk, and I ended up in its actuarial department working with IBM punch cards. Later, while in the Army, I took engineering courses at North Carolina State University and took calculus from R. L. Anderson. After the war, I returned to NC State to finish a degree in electrical engineering and found Andy (Professor Anderson) in the statistics department. My formal statistics education began then and there.

Q: What motivated you and George Box to write Statistics for Experimenters?

A: I had the good fortune to be one of George’s first (of two) graduate students—Sigurd Andersen the other. There was great interest in fractional factorial and response surface designs, and I had many opportunities to give short courses and lectures. This included 32 half-hour TV tapes sponsored by Westinghouse Learning and later, when I was elected to the National Academy of Engineering, I was declared to be the “Johnny Appleseed” of industrial experimental design. When George returned to the USA to head up the Statistical Techniques Research Group at Princeton, it seemed the most logical thing in the world to have a book. Bill Hunter, who was then an undergraduate student at Princeton, was a great help in getting mimeo copies made for a course on design of experiments that both George and I gave in the chemical engineering department. When Bill later went to Wisconsin to work on his PhD, he became a natural added author.


Q: You were the founding editor of Technometrics. What motivated you to establish this journal? Who else worked on the founding of Technometrics?

A: A lot of early conversations at the Gordon Conferences argued the need for a statistics journal for the engineering sciences. Cuthbert Daniel and George were particularly avid advocates. My many short courses for the Chemical Division of the ASQC had fattened its treasury, and launching a technical journal seemed a natural use for this resource. While a graduate student at NC State, I had occasionally helped Sarah Porter and Gertrude Cox proofread and edit early issues of Biometrics, and I just selected myself to be the editor. R. A. Fisher named the journal Technometrics.

Q What accomplishment as president of the ASA did you find most gratifying?

A: I initiated the drive to establish certification for statisticians while president. The motion to begin certification was tabled at the last meeting of my presidential year. Serious family illness forced me to be absent.

William Li, SPES JSM Chair

The Section on Physical and Engineering Sciences (SPES) program for JSM 2015 will feature three invited, three topic-contributed, and five contributed sessions.

Invited

Scaling Up Response Surface Models for Big Geostatistical and Computer Simulation Data, organized by Robert Gramacy of The University of Chicago

Design and Analysis of Mixture Experiments: New Methods with Applications, organized by Scott Cooley of Pacific Northwest National Laboratory

Complex and High-Dimensional Inference in Astrostatistics, organized by Jessi Cisewski of Carnegie Mellon University

Topic-Contributed

Physical Sciences, organized by David Jones of Harvard University

New Developments and Applications in Design of Experiments, organized by Julie Zhou of the University of Victoria

Statistical Innovations in Failure Time Modeling of Complex Systems: Pathway to a Better Decision, organized by Sanjib Basu of Northern Illinois University

Awards for Outstanding Presentations

Michael Crotty, SPES Awards Chair

The section is also pleased to announce the results of its annual competition for contributed papers presented at last year’s JSM in Boston, Massachusetts. The outstanding presentation awards encourage excellence in presentation and help raise the SPES contributed sessions to a higher level.

All awards are based on audience evaluations of each speaker.

Outstanding Presentation Award

Christine Anderson-Cook, Los Alamos National Laboratory: “Identifying the Best 16-Run Regular or Non-Regular Screening Design for 6 to 8 Factors Using Multiple Objectives”

Runner-Up, Outstanding Presentation Award

Hernando Ombao, University of California, Irvine: “Modeling Neuronal Cross-Interactions”

Honorable Mentions

Andrew Robinson, University of Melbourne: “ML vs. MRR: Weibull Parameter Estimation for Making Decisions”

Shan Ba, Procter & Gamble: “Optimal Sliced Latin Hypercube Designs for Computer Experiments with Continuous and Categorical Factors”

Maria Weese, Miami University: “Powerful Supersaturated Designs When Effect Directions Are Known”

Gwendolyn Eadie, McMaster University: “Measuring the Mass of a Galaxy: An Evaluation of the Performance of Bayesian Mass Estimates Using Statistical Simulation”

Winners received a certificate recognizing their accomplishment and a cash award.

The awards for the JSM 2014 best presentations will be presented at the SPES mixer during the 2015 meetings in Seattle. Those who assisted with the evaluations are Po-hsu Chen, Elizabeth Claassen, Stephanie DeHart, Emily Griffith, Laura Lancaster, Rajneesh Rajneesh, and Liz Schiferl.

Visit the section’s website for more news and announcements.

http://r14---sn-aigllney.googlevideo.com/videoplayback?initcwndbps=1120000&ip=2a02%3A2498%3Ae002%3A88%3A59%3A%3A2&key=yt5&signature=C6B672F52E9F377B89628283199DBABD4730AAB4.E03F004C7E25DDA08B9337C13E0A6E68C60E8A97&sparams=dur%2Cid%2Cinitcwndbps%2Cip%2Cipbits%2Citag%2Clmt%2Cmime%2Cmm%2Cmn%2Cms%2Cmv%2Cnh%2Cpl%2Cratebypass%2Csource%2Cupn%2Cexpire&mn=sn-aigllney&mt=1435802827&mv=m&id=o-AMrjelqrLTl_2HFrgTpqQmoM7NMKFmqHCju7RQE0LaX1&ms=au&source=youtube&lmt=1435389068286857&dur=62.972&sver=3&mime=video%2Fmp4&fexp=901816%2C9407141%2C9407146%2C9407573%2C9407662%2C9408142%2C9408420%2C9408710%2C9415077%2C9416126%2C9416215%2C9416655%2C9417191%2C952640%2C966300&itag=22&upn=bQH1Ja7LBek&nh=IgpwcjAyLmxocjE0KgkxMjcuMC4wLjE&pl=32&ipbits=0&ratebypass=yes&mm=31&expire=1435824476&title=Play+Of+The+Day+-+Badminton+Finals+-+Yonex+US+Open+C%27ship+2015

(This article was first published on Data Driven Security, and kindly contributed to R-bloggers)
We are happy to announce that the iptools package is now on CRAN. Formerly only available on GitHub, iptools now compiles under Debian/Ubuntu, Fedora/CentOS/RedHat and Mac OS X (we’re still working on that other operating system).

Oliver (the package co-author and on-CRAN instigator) wrote some excellent vignettes that cover the functionality of the package in-depth, but here’s a short-list of what you can find/expect in iptools:

(This article was first published on Ripples, and kindly contributed to R-bloggers)
Triangles are my favorite shape, three points where two lines meet (Tessellate, Alt-J)

Inspired by recurrence plots and by the Gauss error function, I have done the following plots. The first one represents the recurrence plot of where distance between points is measured by Gauss error function:

Earlier this week a press release from the Linux Foundation formally unveiled The R Consortium: “a group of businesses organized under an open source governance and foundation model to provide support to the R community, the R Foundation and groups and individuals, using, maintaining and distributing R software”.  Mango Solutions were announced as founding silver members alongside the R Foundation; Microsoft and RStudio (Platinum); TIBCO Software Inc. (Gold); and Alteryx, Google, HP, Ketchum Trading and Oracle (Silver).  Clearly we think The R Consortium is idea.  But what does it mean for ordinary R users?

In this paper, we are concerned with two common and related problems for generalized varying-coefficient models, variable selection and constant coefficient identification. Starting with a specification of generalized varying-coefficient models assuming possible nonlinear interactions between the index variable and all other predictors, we propose a polynomial-spline based procedure that simultaneously eliminates irrelevant predictors and identifies predictors that do not interact with the index variable. Our approach is based on a double-penalization strategy where two penalty functions are used for these two related purposes respectively, in a single functional. In a “large p, small n” setting, we demonstrate the convergence rates of the estimator under suitable regularity assumptions. Based on its previous success on parametric models, we use the extended Bayesian information criterion (eBIC) to automatically choose the regularization parameters. Finally, post-penalization estimator is proposed to further reduce the bias of the resulting estimator. Monte Carlo simulations are conducted to examine the finite sample performance of the proposed procedures and an application to a leukemia dataset is presented.

In this paper, we are concerned with two common and related problems for generalized varying-coefficient models, variable selection and constant coefficient identification.

Starting with a specification of generalized varying-coefficient models assuming possible nonlinear interactions between the index variable and all other predictors, we propose a polynomial-spline based procedure that simultaneously eliminates irrelevant predictors and identifies predictors that do not interact with the index variable.

Generalized linear models (GLM) provide an extension of linear models in dealing with different types of responses, including for example binary data and count data  [24]. However, such parametric models are not flexible enough to capture the true underlying relationships between covariates and responses. Of particular interests to us in this paper is the generalized varying-coefficient models (GVCM)  [13] and [3]. Let Y be a response variable and (X,T) is the associated covariates where T is one dimensional and X=(X1,…,Xp)T

The (conditional) mean of the response, μ=E[Y|X,T], takes the form

equation(1)
g(μ)=XTα(T),
Turn MathJax on

where α(T)=(α1(T),…,αp(T))T. The index variable T is usually some variable related to time or age in many applications whose interactions with other predictors are believed to be of importance. Meanwhile, we assume the conditional variance View the MathML source only depends on the conditional mean.

女主角死了，男主角疯了，女主角用尽心思让男主角杀了自己，只为了让高高在上的他掉下来，她最终成功了。（半夜看完某热剧《花**》）

Outlier detection covers the wide range of methods aiming at identifying observations that are considered unusual. Novelty detection, on the other hand, seeks observations among newly generated test data that are exceptional compared with previously observed training data. In many applications, the general existence of novelty is of more interest than identifying the individual novel observations. For instance, in high-throughput cancer treatment screening experiments, it is meaningful to test whether any new treatment effects are seen compared with existing compounds. Here, we present hypothesis tests for such global level novelty. The problem is approached through a set of very general assumptions, making it innovative in relation to the current literature. We introduce test statistics capable of detecting novelty. They operate on local neighborhoods and their null distribution is obtained by the permutation principle. We show that they are valid and able to find different types of novelty, e.g. location and scale alternatives. The performance of the methods is assessed with simulations and with applications to real data sets.

Sample size justification is an important consideration when planning a clinical trial, not only for the main trial but also for any preliminary pilot trial. When the outcome is a continuous variable, the sample size calculation requires an accurate estimate of the standard deviation of the outcome measure. A pilot trial can be used to get an estimate of the standard deviation, which could then be used to anticipate what may be observed in the main trial. However, an important consideration is that pilot trials often estimate the standard deviation parameter imprecisely. This paper looks at how we can choose an external pilot trial sample size in order to minimise the sample size of the overall clinical trial programme, that is, the pilot and the main trial together. We produce a method of calculating the optimal solution to the required pilot trial sample size when the standardised effect size for the main trial is known. However, as it may not be possible to know the standardised effect size to be used prior to the pilot trial, approximate rules are also presented. For a main trial designed with 90% power and two-sided 5% significance, we recommend pilot trial sample sizes per treatment arm of 75, 25, 15 and 10 for standardised effect sizes that are extra small (≤0.1), small (0.2), medium (0.5) or large (0.8), respectively.