-
PR #928 will add GEE for generalized linear models to statsmodels.
currently we are missing some post-estimation results
Stata http://www.ats.ucla.edu/stat/stata/library/gee.htm
R gee http://www.insi…
-
In many real cases the fuzzy sphere actually is composed of polymer chains (for example dendrimers). The original fuzzy sphere does not account for the high q contributions from the fluctuations. Rath…
-
(late night random thought, there are related issues but no time)
proposal/possible API for using prob_weights with survey data, and given that we don't have Stata_like prefix syntax.
define wei…
-
(I ran into this now 3,4 times in the last months for different applications. I might have some experiments from years ago.)
For QMLE or estimating equations or GMM under mis- or incompletely speci…
-
When using `eval` to parse variable's to the power only positive values are considered.
```rust
use kalk::parser;
fn main() {
let mut parser_context = parser::Context::new();
let result…
-
(mainly a reference, and possibly a reference for unit tests)
https://cran.r-project.org/web/packages/drgee/index.html
http://www.degruyter.com/view/j/em.2015.4.issue-1/em-2014-0021/em-2014-0021.xml…
-
related issue: covariance shrinkage, and possible other non-model penalization
PRs, current code in statsmodels
- discrete fit_regularized
- elastic net fit_regularized, GLM, OLS/WLS, minimal R…
-
I'm getting confused about the sign of hessian_factor.
I was adding hessian factor to discrete so that `hessian = sum_i(hessian_factor_i * outer(exog_i))`, in an upcoming PR
However, GLM hessia…
-
A refactoring or enhancement proposal
the outlier test currently uses `get_influence` to get the studentized residuals
GLM and RLM already have the studentized residuals or standardized residuals as…
-
Below there is the traceback from a GEE Binomial fit example (from my notebook comparing GEE, GLM and discrete).
I just discovered that my problem was because I used count data in GEE, instead of the…