High-Performance Open-Source Archive
dim before calling bind_rows() to
account for changes in vctrs (thanks to Davis Vaughan)wald.test() (from aod package) included as internal
function to avoid archivalBug-fix detrend() for handling of vectors in
data
Bug-fix detrend() for character vectors of data
length in id
predicted option in detrend() returns
predicted values
feis() now has option weights to
estimate FEIS models with weighted least squares
Allows estimation of conventional FE models by ‘y ~ x | 1’
Rank deficit columns / NA coefs are now dropped from the model (with warning message)
Bug fix: Omit intercept from transformed data if intercept = TRUE
detrend() allows demeaning by ‘slopes = 1’
Minor documentation and vignette updates
Define collinearity test (R:44:1) dynamic based on QR rank on system
New options se.fit and interval for predict.feis():
Compute standard errors and confidence intervals for predicted
values.
New function detrend(): Detrends the input data by
the predicted values based on the slope parameters within each
group.
Bug fix collinearity handling: now compares equality between predicted x and actual x.
Bug fix for vignette: Add correct suggests version
Minor documentation updates
Bug fix feis(), feistest(),
bsfeistest(): Vcov / Standard errors are now computed even
if some coefficients are NA.
feistest() and bsfeistest(): added
terms option to perform Hausman / Chi_sq test on subset of
coefficients only.
feis(): added tol option for detecting
linear dependencies in slopes.
feis(): available S3 methods extended by coef,
deviance, df.residual, fitted, formula, hatvalues, model.matrix, nobs,
predict, residuals, sigma, terms, vcov.
Added general info page feisr-package.
Included dplyr bind_rows() for more efficient
transformation of large data (depends on dplyr (>= 1.0.0)). Old
fall-back code available for dplyr < 1.0.0.
Added vignette for package use.
More efficient use of solve / crossprod
Manual updated, minor corrections.
Added citation for package.
Corrected linkage and definition of S3 methods.
Need mirroring services?
Contact our team at info@vpspulse.com.
Mirror powered by VPSpulse
Infrastructure sponsored by VPSPulse & Secure Payments by ArionPay.