statsmodels.sandbox.regression.gmm.NonlinearIVGMM.fitgmm NonlinearIVGMM.fitgmm(start, weights=None, optim_method='bfgs', optim_args=None) estimate parameters using GMM Parameters: start : array_like starting values for minimization weights : array weighting matrix for moment conditions. If weights is None, then the identity matrix is used Returns: paramest : array estimated parameters Notes todo: add fixed parameter option, not here ??? uses scipy.optimize.fmin

statsmodels.graphics.functional.rainbowplot statsmodels.graphics.functional.rainbowplot(data, xdata=None, depth=None, method='MBD', ax=None, cmap=None) [source] Create a rainbow plot for a set of curves. A rainbow plot contains line plots of all curves in the dataset, colored in order of functional depth. The median curve is shown in black. Parameters: data : sequence of ndarrays or 2-D ndarray The vectors of functions to create a functional boxplot from. If a sequence of 1-D arrays, these

statsmodels.duration.hazard_regression.PHRegResults.predict PHRegResults.predict(endog=None, exog=None, strata=None, offset=None, pred_type='lhr') [source] Returns predicted values from the fitted proportional hazards regression model. Parameters: params : array-;like The proportional hazards model parameters. endog : array-like Duration (time) values at which the predictions are made. Only used if pred_type is either ?cumhaz? or ?surv?. If using model exog, defaults to model endog (time

statsmodels.discrete.discrete_model.Poisson.score Poisson.score(params) [source] Poisson model score (gradient) vector of the log-likelihood Parameters: params : array-like The parameters of the model Returns: score : ndarray, 1-D The score vector of the model, i.e. the first derivative of the loglikelihood function, evaluated at params Notes where the loglinear model is assumed

statsmodels.genmod.families.links.NegativeBinomial.deriv NegativeBinomial.deriv(p) [source] Derivative of the negative binomial transform Parameters: p : array-like Mean parameters Returns: g?(p) : array The derivative of the negative binomial transform link function Notes g?(x) = 1/(x+alpha*x^2)

statsmodels.sandbox.tsa.fftarma.ArmaFft.impulse_response ArmaFft.impulse_response(nobs=None) get the impulse response function (MA representation) for ARMA process Parameters: ma : array_like, 1d moving average lag polynomial ar : array_like, 1d auto regressive lag polynomial nobs : int number of observations to calculate Returns: ir : array, 1d impulse response function with nobs elements Notes This is the same as finding the MA representation of an ARMA(p,q). By reversing the r

statsmodels.genmod.families.family.Poisson.predict Poisson.predict(mu) Linear predictors based on given mu values. Parameters: mu : array The mean response variables Returns: lin_pred : array Linear predictors based on the mean response variables. The value of the link function at the given mu.

statsmodels.regression.linear_model.GLSAR.hessian GLSAR.hessian(params) The Hessian matrix of the model

statsmodels.nonparametric.kernel_density.KDEMultivariateConditional.cdf KDEMultivariateConditional.cdf(endog_predict=None, exog_predict=None) [source] Cumulative distribution function for the conditional density. Parameters: endog_predict: array_like, optional : The evaluation dependent variables at which the cdf is estimated. If not specified the training dependent variables are used. exog_predict: array_like, optional : The evaluation independent variables at which the cdf is estimated

statsmodels.sandbox.distributions.extras.SkewNorm_gen.nnlf SkewNorm_gen.nnlf(theta, x) Return negative loglikelihood function Notes This is -sum(log pdf(x, theta), axis=0) where theta are the parameters (including loc and scale).