statsmodels.robust.norms.TukeyBiweight class statsmodels.robust.norms.TukeyBiweight(c=4.685) [source] Tukey?s biweight function for M-estimation. Parameters: c : float, optional The tuning constant for Tukey?s Biweight. The default value is c = 4.685. Notes Tukey?s biweight is sometime?s called bisquare. Methods psi(z) The psi function for Tukey?s biweight estimator psi_deriv(z) The derivative of Tukey?s biweight psi function rho(z) The robust criterion function for Tukey?s biweight e

statsmodels.genmod.generalized_estimating_equations.GEEResults.fittedvalues static GEEResults.fittedvalues() [source] Returns the fitted values from the model.

statsmodels.stats.sandwich_covariance.cov_cluster_2groups statsmodels.stats.sandwich_covariance.cov_cluster_2groups(results, group, group2=None, use_correction=True) [source] cluster robust covariance matrix for two groups/clusters Parameters: results : result instance result of a regression, uses results.model.exog and results.resid TODO: this should use wexog instead use_correction : bool If true (default), then the small sample correction factor is used. Returns: cov_both : ndarray

statsmodels.tsa.vector_ar.var_model.VAR.select_order VAR.select_order(maxlags=None, verbose=True) [source] Compute lag order selections based on each of the available information criteria Parameters: maxlags : int if None, defaults to 12 * (nobs/100.)**(1./4) verbose : bool, default True If True, print table of info criteria and selected orders Returns: selections : dict {info_crit -> selected_order}

statsmodels.sandbox.regression.gmm.IVGMMResults.compare_j IVGMMResults.compare_j(other) overidentification test for comparing two nested gmm estimates This assumes that some moment restrictions have been dropped in one of the GMM estimates relative to the other. Not tested yet We are comparing two separately estimated models, that use different weighting matrices. It is not guaranteed that the resulting difference is positive. TODO: Check in which cases Stata programs use the same weigths

statsmodels.duration.hazard_regression.PHReg.breslow_gradient PHReg.breslow_gradient(params) [source] Returns the gradient of the log partial likelihood, using the Breslow method to handle tied times.

statsmodels.stats.power.GofChisquarePower.plot_power GofChisquarePower.plot_power(dep_var='nobs', nobs=None, effect_size=None, alpha=0.05, ax=None, title=None, plt_kwds=None, **kwds) plot power with number of observations or effect size on x-axis Parameters: dep_var : string in [?nobs?, ?effect_size?, ?alpha?] This specifies which variable is used for the horizontal axis. If dep_var=?nobs? (default), then one curve is created for each value of effect_size. If dep_var=?effect_size? or alpha

statsmodels.genmod.families.links.NegativeBinomial.inverse NegativeBinomial.inverse(z) [source] Inverse of the negative binomial transform Parameters: z : array-like The value of the inverse of the negative binomial link at p. Returns : ??- : p : array Mean parameters Notes g^(-1)(z) = exp(z)/(alpha*(1-exp(z)))

statsmodels.genmod.generalized_linear_model.GLMResults.resid_anscombe static GLMResults.resid_anscombe() [source]

statsmodels.tsa.ar_model.ARResults.cov_params ARResults.cov_params(r_matrix=None, column=None, scale=None, cov_p=None, other=None) Returns the variance/covariance matrix. The variance/covariance matrix can be of a linear contrast of the estimates of params or all params multiplied by scale which will usually be an estimate of sigma^2. Scale is assumed to be a scalar. Parameters: r_matrix : array-like Can be 1d, or 2d. Can be used alone or with other. column : array-like, optional Must be