7,391 reputation
1633
bio website thinkinator.com
location Washington DC, United States
age
visits member for 3 years, 10 months
seen 2 days ago

Long-time R user, new Stata user, occasional Stan user, and currently learning Scala. Data Scientist at ERI, an incredible company where we do everything from fraud detection and visualization, to training and and text mining.


Aug
26
awarded  Enlightened
Aug
26
awarded  Nice Answer
Jul
25
answered How to tell which recipe rating is better?
Jul
25
comment Specifying lag in `dlnm` when passing arguments to `crossbasis`
The $y=x1*x2$ formula is shorthand for $y=x1+x2+x1:x2$. That is, it introduces the interaction between the two variables. Not sure if you're confusing autocorrelation and correlation: autocorrelation is correlation of a variable with lagged versions of itself. ("Auto" = "Self")
Jul
14
comment Best way to turn a date into a numerical feature?
Why don't you want to calculate the difference between dates? You can also include an age^2 feature or even a unit less number like the age that year divided by the median age that year. Binning ages into intervals will throw away information and us usually not a good idea.
Jul
12
comment Multiplicative errors for linear model
Could you instead change the original formulation to: $Y_i=(ax_i+b)(e^{\epsilon_i})$, where $\epsilon=N(0,\sigma^2)$? Or $Y_i=(ax_i+b)(1+\epsilon_i)$?
Jul
12
comment Multiplicative errors for linear model
You're using a Gaussian error with mean 1 because you want it to be a ratio? You can hope that a small enough variance would keep it from going negative, but you really need something like a Gamma distribution for that. Unfortunately, I don't know enough to venture farther.
Jul
6
answered Best way to turn a date into a numerical feature?
Jul
2
awarded  Curious
Jun
7
comment Difference between SUR and SEM
I've removed the "sem" tag, which refers to Structural Equation Models, which is different from Simultaneous Equation Models. (EDIT: Not an expert in SUR, but my first impression is that SUR is a subset of simultaneous equation models.)
Jun
7
revised Difference between SUR and SEM
edited tags
Jun
2
comment Similarity of time series
Again, proof by induction. Base case is a window of width 1, i.e. the original points for a difference of zero. If you have a window width of N and add a point on each side for a window of width N+2, how much can those two points change your average? Well, calculate the greatest difference between any two points in the time series, let's call it D. Etc...
Jun
1
answered Similarity of time series
May
29
comment Is it possible to change a hypothesis to match observed data (aka fishing expedition) and avoid an increase in Type I errors?
@post-hoc: If you use the holdout as a follow-on validation of your best theory -- as suggested by Peter Flom -- that's fine. But if you dip into the non-holdout data, then bang a few hypothesis against the holdout, then dip back into the non-holdout data and repeat, it won't be accomplishing what you think.
May
27
comment Is it possible to change a hypothesis to match observed data (aka fishing expedition) and avoid an increase in Type I errors?
Taking Frank Harrell's comment a step further: it's legitimate to explore some data and publish an intriguing finding... as an intriguing, exploratory finding that's subject to being reproduced/validated. The downside is: if someone else confirms your findings they may well get the glory, and if others do not confirm your results you were fooled by a spurious correlation. Bad if you have a big ego. Not to mention you would need to make your data and procedures publicly available, which many practitioners in many fields wont't do. And you should follow up with new data rather than moving on.
Apr
18
awarded  Good Answer
Apr
17
answered What will be the process of statistical analysis with Survey data?
Apr
17
comment What does “fiducial” mean (in the context of statistics)?
+1 As far as I understand it, Fisher was trying to plant the Bayesian credible interval in his frequentist garden. (He famously despised the Bayesian approach and I believe even coined the term "Bayesian" as an expression of his contempt.)
Apr
17
awarded  Mortarboard
Apr
17
awarded  Nice Answer