meta user
network profile
|
May 17 |
awarded | Promoter |
|
May 15 |
asked | Bootstrapping power estimates for a bootstrap test |
|
May 3 |
comment |
Is the p-postulate true? What do you understand "evidence" to be here? Do you have an exact definition in mind? |
|
Apr 19 |
comment |
Can I use a Z-score with skewed and non-normal data? I agree with @MaartenBuis but unlike him I will downvote this. |
|
Apr 17 |
answered | Random effects are not normal |
|
Apr 16 |
comment |
ANOVA: dump data points instead of graphs First off, this seems to be a R language only question which is not directly statistics related and therefore off-topic. The question is also not specific enough - what kind of plot does R create for you and you want to have points for? |
|
Apr 11 |
comment |
What should I do when a confidence interval includes an impossible range of values? It's pretty impossible. Let's assume that your sample size is large enough and you are approximately normally distributed. Then your confidence interval is your sample mean +/- 2*Std.Error. The Standard error is sample sd/sqrt(n). This should be (decent sample size, right?) much smaller than the sample sd. Let's say it's 20% of your sd. This means that if all your assumptions are true are considerable percentage (>30%) of your observations would have to be > 30. Since they aren't (and can't be) your assumption are wrong. |
|
Apr 11 |
awarded | Yearling |
|
Apr 8 |
comment |
Can I justify performing a two-way anova where data is normally distributed but has heterogeneous variances? There is not enough information available here. How large is your data? Is it balanced? How large is the ratio of variances? If your data is unbalanced, does the larger group have larger or smaller variance? Generally speaking, balanced ANOVA is rather robust against heteroscedascity and the most problematic case is if the smaller group has larger variance. |
|
Apr 5 |
answered | Is it possible to compare two feature selections algorithms by cross-validations? |
|
Mar 21 |
awarded | Nice Answer |
|
Mar 21 |
comment |
Error in JAGS code I think Glen is entirely correct and the problem is your definition of denom. You either need to spell it out since you have just 3 elements anyway to exp(p[1])+exp(p[2])+exp(p[3]) or as a more general solution use a loop to define a additional vector like exponential.of.p[i] <- exp(p[i]) and the sum over that. |
|
Mar 21 |
answered | p-values of Mann-Whitney U test identical for raw and log-transformed data |
|
Mar 19 |
comment |
Finding patterns in data @AntoineVernet This is often a good way but it depends on the first two components explaining a substantial share of the total variation. |
|
Mar 19 |
comment |
Queries on the Bayesian method With regards to the 2nd question, there are advantages (and in some cases and applications disadvantages) to a bayesian approach beyond using informative priors. For example, bayesian models works very well if you have to average over hierarchical structures. Usually you don't switch to frequentist methods with more data, you just change to vague and hopefully noninformative priors. |
|
Mar 19 |
comment |
Queries on the Bayesian method Frankly, I still do not understand what is going on regarding questions 1 and 3. What does for example"There are parameters of the model which has captured a very wide range - say, 10% to 90%. It does not give me comfort, rather, it may show that the expert panel inputs missed out on the clear range" mean? Do you any clear conflict between the prior and data? |
|
Mar 18 |
answered | Automated Normality Test on large samples |
|
Mar 15 |
comment |
Getting started with biclustering @chl The first link to the Pardalos slides seems to be dead. Does anyone know of an alternative location? |
|
Mar 8 |
comment |
p-values for p-values? This is a bare link and does not describe the bare connection to the question. Though it is good reading. |
|
Mar 1 |
answered | Avoiding Type I error in multipe t-tests of different groups of data |
