csgillespie

6,094 reputation

22556

bio	website	mas.ncl.ac.uk/~ncsg3
	location	Newcastle, United Kingdom
	age	34
visits	member for	2 years, 10 months
	seen	10 hours ago
stats	profile views	912

I'm a statistics lecturer at Newcastle University, UK (Uni homepage). My research interests are parallel computing, Bayesian statistics and stochastic kinetic models.

I also

run some R courses at Newcastle University
occasionally write the odd blog post
run on site R training

	6,094 reputation	bio	website	mas.ncl.ac.uk/~ncsg3	visits	member for	2 years, 10 months
22556 badges		location	Newcastle, United Kingdom		seen	10 hours ago

summary answers questions tags badges favorites bounties reputation activity

167 Comments

suggestions reviews revisions comments badges posts accepts all

Feb 25	comment	Fitting models to empirical data: do fits for random samples converge on the true fit? As a matter of interest - how big? Sample size, range of data.
Jan 23	comment	Reversible jump MCMC code (Matlab or R) Yes, I quite liked the King book. I've never read the WinBUGS book.
Jan 23	comment	Reversible jump MCMC code (Matlab or R) @Tomas "in anger" == "a real problem". Oh, my comment should say "never needed to use"
Jan 23	comment	Reversible jump MCMC code (Matlab or R) @Tomas The book uses RJMCMC for model choice. I found the description quite nice. I've needed to use RJMCMC in anger though.
Jan 22	comment	Reversible jump MCMC code (Matlab or R) @Tomas Someone edited my answer and changed R to WinBUGS. From what I recall, the book provides R code.
Apr 23	comment	How to reduce autocorrelation in Metropolis algorithm? "...reaching the memory limit on my computer to use" - I presume that you are thinning as you go, rather than post processing your samples?
Dec 20	comment	Model prediction for every possible outcome That means that age can take infinitely values: 18, 18.00001, 18.00002,.... plus lots of values inbetween
Dec 20	comment	Model prediction for every possible outcome Surely age will be a CTN variable and not a factor?
Oct 5	comment	What is a good general purpose plain text data format like that used for Bibtex? I would strongly advise against using XML. XML is meant to be edited by a software tool, not by hand. Let me repeat - you shouldn't really edit XML by hand. Just because you can, doesn't mean you should. Just imagine, every time you want to edit a node, you call up `xmlTreeParse`, followed by an additional six commands to change a single value. Keep the file format simple. It will save you grief in the future.
Sep 19	comment	Finding a function for data points with disturbances @whuber: The first figure clearly shows the P and T depend on time. However, predicting P with just T will get you close to your goal, i.e. if you know T, you will be able to predict P fairly well. By having additional terms in the model, I hoped to mope up the remaining variation.
Jul 25	comment	How to find groupings (trajectories) among longitudinal data? @Andy: Good point. I've included a quick simulation. Everything seems OK, but there might be a more optimal solution.
Jul 8	comment	Why is R plotting standardized residuals against theoretical quantiles in a Q-Q plot? @Mark: Just google "standardised residuals" and look at the first few hits.
Jul 5	comment	Multiple simulations of a system under different conditions - paired data? But why do you want to do a t-test. Just simulate the process N times and calculate the difference. A better method would be to simulate the difference over time, and plot the difference with associated error
Jul 4	comment	How can I create a Graph of a Probability Density Function from Percentiles? You have the probability from the percentiles. What's the probability of getting over the 50th value - 0.5. What's the probability of getting over the 95th value, 0.05.
Jun 27	comment	How can I create a Graph of a Probability Density Function from Percentiles? Yes. If players start at 1500, that implies that people are either good or bad, with very few people in the middle - hence the U shape.
Jun 27	comment	How can I create a Graph of a Probability Density Function from Percentiles? Would you include some example data?
Jun 16	comment	If I count cells should I consider them as Poisson distributed? @stan: Once you calculate the ratio: Lph*/Lph, your data is no longer discrete - or am I missing something?
May 5	comment	How to use rejection sampling to generate draws from Unit Exponential One small thing, rather than constantly resizing your vector, start off with: `x = numeric(100000)` This will be much faster.
Mar 1	comment	Algebraic definition of a residual from a regression I presume this is homework? What have you tried so far?
Feb 22	comment	Power analysis for survival analysis I hope you don't mind, but I converted your question into latex. One thing, should $Z_{\alpha}$ not be $Z_{\alpha/2}$