Josef Perktold

Results 172 issues of Josef Perktold

PASS has power and sample size computation for cluster randomized trials, I never looked at the details Song and Ahn looks like a good starting point, although it has only...

type-enh
comp-genmod
comp-discrete
comp-stats

(just parking some references, I only read parts of abstracts, list collected by citations and related articles) This sounds like the weighted ATE equivalent of get_prediction with weights and average...

type-enh
comp-treatment

based on discussion in #8380 loglike score and hessian in MNLogit seem to work for fractional data. This needs to be verified with some examples and results need to be...

type-enh
comp-discrete

(parking another helper function, so it doesn't get lost) Written to understand orthogonal complement which is heavily used in VECM/cointegration. Should also show up in canonical correlation, reduced rank regression...

comp-tsa
type-enh
comp-tools

two or k-sample comparison with unequal distributions, like Behrens-Fisher problem if variance differs. If we have two samples with different distributions, variance, skewness and kurtosis, then comparing means, medians or...

type-enh
comp-stats

Currently we don't have power and sample size functions for statsmodels.stats.nonparametric.RankCompareResult I just found this reference, but didn't specifically search for this topic Happ, Martin, Arne C. Bathke, and Edgar...

type-enh
comp-stats

Currently test_ordinal_association only allows for two-sided trend test. related issue #1130 k-sample trend tests instead we often want to test also for one-sided alternatives, e.g. positive trend in dose-response models...

type-enh
comp-stats

k-sample test with Null hypthesis that mean or median are the same and alternative hypothesis that the mean or median are ordered, increasing or decreasing. see https://groups.google.com/d/msg/pystatsmodels/PLD-QdF7OxI/9UI982MbsBoJ Cuzick's trend test...

type-enh
comp-stats

(mainly parking idea and references and links) Stata has suest command to provide inference for combining tw or more models Wooldridge mentioned in a tweet that suest cannot be used...

type-enh
comp-genmod
comp-discrete
comp-regression

general issue: I have no idea how well this works, and AFAIK we don't have an example How well does patsy handle this? What is the parameterization? Which collinear columns...

type-enh
comp-base
design