14,047 reputation
42351
bio website
location England
age
visits member for 2 years, 4 months
seen 10 mins ago

1h
comment Still confused with the p-value definition
@HemantRupani: Yes, really: the quote is referencing Fisher's disjunction.
1h
answered Still confused with the p-value definition
1h
revised Modelling proportional hazards in Cox Model using coxph in R
fixed typo, improved formatting
2h
comment How to perform a user defined hypothesis testing on regression coefficient uisng R?
Please be polite! See here for the explanation, or e.g. Faraway (2002), Practical Regression & ANOVA using R, Ch.3 for the explanation plus R code. How to use specific statistical software isn't on-topic here.
4h
comment Is it correct to use plus or minus symbol before standard deviation?
$\bar{x} \pm 2\mathrm{SD}$ could be an asymptotic 95% prediction interval for a Gaussian random variable (asymptotic because the sample mean, & perhaps the standard deviation, are estimated from the data; so exact prediction intervals are somewhat wider).
5h
comment Modelling proportional hazards in Cox Model using coxph in R
So are you asking how to estimate the intercept that's been applied to a known baseline survival function? (If so please edit the question to reflect this.)
21h
revised Creating an interaction term with 2 continuous variables: What to do?
added 224 characters in body
22h
answered Creating an interaction term with 2 continuous variables: What to do?
22h
reviewed Approve Show that $\int_0^t X(t,s) dB(s)$ is a Gaussian random variable $Y(t)$
23h
comment Modelling proportional hazards in Cox Model using coxph in R
I'm a bit confused because you seem to understand why the intercept isn't estimable (i.e. any estimate would fit the data as well as any other) - what would you do with it? Are you asking how to estimate the baseline survival function $S_0$?
1d
comment Logit model - none of the cases vary on one predictor
Welcome to Cross Validated! The name for the situation is "separation", & there's no reason to say the logit model can't handle it. Maximum-likelihood gives infinite odds ratio estimates, which might well be all right; for some other methods see here.
1d
comment Standard multiple regression and multiple dependent variables. and bonferroni correction?
What would the three multiple regressions for one dependent variable be then?
Apr
16
comment Do I drop insignificant parameters from a model? Should I use stepwise regression?
Welcome to Cross Validated! There are already good answers to these questions on our site: see here, here, here, here, & here. Please have a look & let us know if there's anything you still want to know.
Apr
16
comment Trust of coefficients of Logistic Regression
Why do you say "the standard errors of the coefficients does not necessarily seems to incorporate the information on how rare a particular feature is"?
Apr
16
reviewed Close Are categorical variables standardized differently in penalized regression?
Apr
16
reviewed Leave Open Inequalities for maxima over random functions
Apr
16
comment formulating questionnaires
Welcome to Cross Validated! There are whole books on survey methodology, so please edit your question to address one specific aspect that's giving you difficulty, & explain why.
Apr
16
revised Inequalities for maxima over random functions
formatted title (at first I thought Maxima meant the software)
Apr
15
reviewed Leave Open Scaling factor in R?
Apr
15
reviewed Leave Open How to remove horizontal bar in Hindi word Matlab