Reputation
390
Top tag
Next privilege 500 Rep.
Access review queues
Badges
2 11
Newest
 Enthusiast
Impact
~5k people reached

  • 0 posts edited
  • 0 helpful flags
  • 14 votes cast
May
23
comment the test statistic of the sum of a multiple regression
$var(a+b)=var(a)+var(b)+2cov(a,b)$. You're missing the latter term. Otherwise, I'm not sure what the question is?
May
21
answered If $\ln(y) = 5 - 0.1X $what is the elasticity of $Y$ with respect to $X$, when $X=10$?
May
20
answered Class or type variables as features in polynomial regression algrorithm
May
19
comment Confidence Interval for Nonlinear Regression using F-Test - lmfit
Well, my guess came from the F-statistics for the standard OLS regression.(see, for example, here: pages.uoregon.edu/aarong/teaching/G4075_Outline/node4.html). So, if anything, it seems like the formula from your link is a more generalized version of that. Whether it can be generalized even further, I do not know.
May
19
comment Confidence Interval for Nonlinear Regression using F-Test - lmfit
My guess would be they represent the sum of square residuals.
May
12
answered Demeaned fixed effects invariant to base category
May
4
comment Approximation technique when data is missing?
I don't follow your second paragraph. For example MAR literally stands for missing-at-random, which would imply no correlation whatsoever to the dependent variable. This would seem to meet your needs in paragraph 3. Also, if your missing data is really independent of all other variables, then I don't see how you can do any imputation. Imputations are done by assuming a relationship between independent variables and using the other values to "guess" the missing values.
Apr
22
awarded  Enthusiast
Apr
21
comment How to curve fit an unknown function?
You should look into non-parametric regressions like kernel regressions or local linear or series regressions.
Apr
15
comment Comparison of parameter: two different populations
Perhaps then some more information on the problem and the regression(s) you want to run
Apr
15
comment Comparison of parameter: two different populations
From your setup, that looks like 2 separate regressions, not two parameters in one regression.
Apr
13
awarded  Benefactor
Apr
13
accepted Bayesian Updating with 1 Signal but 2 Unknowns
Apr
12
comment Bayesian Updating with 1 Signal but 2 Unknowns
I'm a little confused. Doesn't $\mu_1$ depend on $\alpha_1$ and the updated value of $p_1$? If so then how can we solve your equation straight up. Wouldn't we need to find a "fixed point" so that the equation is satisfied? (That is marginal value for $\alpha$ on the RHS that is consistent with the joint distribution on the LHS).
Apr
10
answered How to report significant digits in coefficient of determination?
Apr
8
answered Differentiate intergral function
Apr
8
revised Bayesian Updating with 1 Signal but 2 Unknowns
added 13 characters in body
Apr
8
comment Bayesian Updating with 1 Signal but 2 Unknowns
@DanielWeissman Good Catch, Fixing now
Apr
6
comment How do I get the proper p-value of a time series average of regression coefficients?
Is it possible to bootstrap? en.wikipedia.org/wiki/Bootstrapping_(statistics)
Apr
6
answered every pdf can be regarded as a marginal distribution of a joint pdf