7,876 reputation
1033
bio website
location San Jose, CA
age 34
visits member for 2 years, 11 months
seen 3 hours ago

Mad data scientist.


1d
comment Interpretation of log transformed predictor in negative binomial regression
And what problem does that cause with the NBR?
1d
comment Bias in jury selection?
@JonathanMarch Unfortunately, I am not sure I know how to do a joint one-sided test.
1d
revised Bias in jury selection?
added 1 character in body
1d
revised Bias in jury selection?
added 36 characters in body
1d
revised Bias in jury selection?
added 1729 characters in body
1d
comment Bias in jury selection?
@whuber My ideas was to use group B to estimate the counterfactual, bias-free rate of rejection and use it as a yardstick for the other 3 groups. A and D should be less likely to be rejected and C should be more, given the story.
2d
comment Distances for binary and non binary categorical data
There do seem to be several possibilities: www-users.cs.umn.edu/~sboriah/PDFs/BoriahBCK2008.pdf
2d
comment Interpretation of log transformed predictor in negative binomial regression
Why did you transform the predictor?
Sep
17
comment How to best make millions of forecasts using time series data?
This is a nice start: robjhyndman.com/talks/automaticforecasting
Sep
17
comment Bias in jury selection?
I am not sure I understand (1). For (2), I get very similar results with a logit model, which makes different distributional assumptions, so there is some robustness. There's not enough data to do something less parametric, though that may be my own ignorance in this area.
Sep
17
comment Distances for binary and non binary categorical data
I don't know any way to do this other than turning it into a set of binary indicators.
Sep
17
answered Bias in jury selection?
Sep
17
comment Homework problem. What is the answer supposed to look like?
Here's a hint. X is random variable with known properties. You are asked what is the probability that X exceeds 203. Intuitively, X tends to hover around 200, so the probability that X is slightly greater than the mean of 200 should be a bit less than one half.
Sep
17
revised Homework problem. What is the answer supposed to look like?
edited tags
Sep
16
comment test of assocation for continuous IV, binary DV
Some good ideas here: ats.ucla.edu/stat/mult_pkg/whatstat
Sep
16
revised Adjust for everything you have in propensity score?
added 257 characters in body
Sep
16
comment Back transforming regression results when modeling log(y)
@Glen Do a search for Duan smearing on this site.
Sep
15
revised How useful is Minitab in the real world?
deleted 189 characters in body
Sep
15
revised How useful is Minitab in the real world?
added 164 characters in body
Sep
15
revised How useful is Minitab in the real world?
added 164 characters in body