David Andrews

Note: this is a picture of R. A. Fisher, not D.F.A.

Draft Ph.D. Comprehensive Examination questions by DFA

 

Symbolic Computation

March 2004. An R implementation of Symbolic Computation for Statistical Inferrence permits the calculation of expressions for moments cumulants asymptotic expansions etc. in R. The expressions may then be applied to data and numerically evaluated directly.

PureStats is a symbolic calculator applet for symbolic computation for statistical inference. It currently handles moments, cumulants, deviances, their expected values, k-statistics and much more. These are the computations discussed in chapters 3-5 of Symbolic Computation for Statistical Inference to be published this summer. Implementation of bootstrap, Edgeworth, and saddlepoint methods are underway.

The applet requires a Java2 enabled browser.

Diversions

Monty Hall A Caffeine free simulation
Studet's n-free t-statistic Student's t-statistic uncluttered.

Research

Student's s-statistic A quantile stable statistic. Presented ISI (1999) Helsinki.

Asymptotic Expansions of Moments and Cumulants A postscript version of the paper presented at the Workshop on Symbolic Computation in Statistics, Montreal September 1997, Statistics and Computing 2001. This presents basic procedures for symbolic computation for statistical inference with examples including those of symbolic bootstrap calculations.

Asymptotic Expansions of Moments and Cumulants A Mathematica notebook accompanying the paper presented at the Workshop on Symbolic Computation in Statistics, Montreal, September 1997.

Edgeworth, Cornish Fisher and Saddlepoint Expansions A Mathematica notebook for computing expansions of the above. This is in a preminary form and is provided here on a need to know basis.

Signed Roots of Likelihood Ratios This paper presents an elementary method for computing expansions of signed roots of log likelihood ratios. These are used to study the robustness of likelihood methods.

Signed roots of likelihood ratios A Mathematica notebook accompanying the above paper.

Quantile Stabilizing transformations This paper procedures for computing a quantile stabilizing transformation to produce BCA-like confidence intervals in closed form.

Data from Andrews and Herzberg (ftp) For educational and research purposes only.

Operator Calculus for Bootstrap Calculations A Mathematica Notebook associated with a talk presented Chicago, August 1996

Toys