docs/source/release/version0.5.rst
statsmodels 0.5 is a large and very exciting release that brings together a year of work done by 38 authors, including over 2000 commits. It contains many new features and a large amount of bug fixes detailed below.
See the :ref:list of fixed issues <issues_list_05> for specific closed issues.
The following major new features appear in this version.
statsmodels now supports fitting models with a formula. This functionality is provided by patsy <https://patsy.readthedocs.org/en/latest/>_. Patsy is now a dependency for statsmodels. Models can be individually imported from the statsmodels.formula.api namespace or you can import them all as::
import statsmodels.formula.api as smf
Alternatively, each model in the usual statsmodels.api namespace has a from_formula classmethod that will create a model using a formula. Formulas are also available for specifying linear hypothesis tests using the t_test and f_test methods after model fitting. A typical workflow can now look something like this.
.. code-block:: python
import numpy as np
import pandas as pd
import statsmodels.formula.api as smf
url = 'https://raw.githubusercontent.com/vincentarelbundock/Rdatasets/csv/HistData/Guerry.csv'
data = pd.read_csv(url)
# Fit regression model (using the natural log of one of the regressors)
results = smf.ols('Lottery ~ Literacy + np.log(Pop1831)', data=data).fit()
See :ref:here for some more documentation of using formulas in statsmodels <formula_examples>
Empirical Likelihood-Based Inference for moments of univariate and multivariate variables is available as well as EL-based ANOVA tests. EL-based linear regression, including the regression through the origin model. In addition, the accelerated failure time model for inference on a linear regression model with a randomly right censored endogenous variable is available.
Support for ANOVA is now available including type I, II, and III sums of squares. See :ref:anova.
.. currentmodule:: statsmodels.nonparametric
Kernel density estimation has been extended to handle multivariate estimation as well via product kernels. It is available as :class:sm.nonparametric.KDEMultivariate <kernel_density.KDEMultivariate>. It supports least squares and maximum likelihood cross-validation for bandwidth estimation, as well as mixed continuous, ordered, and unordered categorical data. Conditional density estimation is also available via :class:sm.nonparametric.KDEMUltivariateConditional <kernel_density.KDEMultivariateConditional>.
Kernel regression models are now available via :class:sm.nonparametric.KernelReg <kernel_regression.KernelReg>. It is based on the product kernel mentioned above, so it also has the same set of features including support for cross-validation as well as support for estimation mixed continuous and categorical variables. Censored kernel regression is also provided by kernel_regression.KernelCensoredReg.
.. currentmodule:: statsmodels.regression.quantile_regression
Quantile regression is supported via the :class:sm.QuantReg <QuantReg> class. Kernel and bandwidth selection options are available for estimating the asymptotic covariance matrix using a kernel density estimator.
.. currentmodule:: statsmodels.discrete.discrete_model
It is now possible to fit negative binomial models for count data via maximum-likelihood using the :class:sm.NegativeBinomial <NegativeBinomial> class. NB1, NB2, and geometric variance specifications are available.
A new optimization method has been added to the discrete models, which includes Logit, Probit, MNLogit and Poisson, that makes it possible to estimate the models with an l1, linear, penalization. This shrinks parameters towards zero and can set parameters that are not very different from zero to zero. This is especially useful if there are a large number of explanatory variables and a large associated number of parameters. CVXOPT <https://cvxopt.org/>_ is now an optional dependency that can be used for fitting these models.
.. currentmodule:: statsmodels.graphics
ProbPlot object has been added to provide a simple interface to create P-P, Q-Q, and probability plots with options to fit a distribution and show various reference lines. In the case of Q-Q and P-P plots, two different samples can be compared with the other keyword argument. :func:sm.graphics.ProbPlot <gofplots.ProbPlot>.. code-block:: python
import numpy as np import statsmodels.api as sm x = np.random.normal(loc=1.12, scale=0.25, size=37) y = np.random.normal(loc=0.75, scale=0.45, size=37) ppx = sm.ProbPlot(x) ppy = sm.ProbPlot(y) fig1 = ppx.qqplot() fig2 = ppx.qqplot(other=ppy)
Mosaic Plot: Create a mosaic plot from a contingency table. This allows you to visualize multivariate categorical data in a rigorous and informative way. Available with :func:sm.graphics.mosaic <mosaicplot.mosaic>.
Interaction Plot: Interaction plots now handle categorical factors as well as other improvements. :func:sm.graphics.interaction_plot <factorplots.interaction_plot>.
Regression Plots: The regression plots have been refactored and improved. They can now handle pandas objects and regression results instances appropriately. See :func:sm.graphics.plot_fit <regressionplots.plot_fit>, :func:sm.graphics.plot_regress_exog <regressionplots.plot_regress_exog>, :func:sm.graphics.plot_partregress <regressionplots.plot_partregress>, :func:sm.graphics.plot_ccpr <regressionplots.plot_ccpr>, :func:sm.graphics.abline_plot <regressionplots.abline_plot>, :func:sm.graphics.influence_plot <regressionplots.influence_plot>, and :func:sm.graphics.plot_leverage_resid2 <regressionplots.plot_leverage_resid2>.
.. currentmodule:: statsmodels.stats.power
The power module (statsmodels.stats.power) currently implements power and sample size calculations for the t-tests (:class:sm.stats.TTestPower <TTestPower>, :class:sm.stats.TTestIndPower <TTestIndPower>), normal based test (:class:sm.stats.NormIndPower <NormIndPower>), F-tests (:class:sm.stats.FTestPower <FTestPower>, :class:sm.stats.FTestAnovaPower <FTestAnovaPower>) and Chisquare goodness of fit (:class:sm.stats.GofChisquarePower <GofChisquarePower>) test. The implementation is class based, but the module also provides three shortcut functions, :func:sm.stats.tt_solve_power <tt_solve_power>, :func:sm.stats.tt_ind_solve_power <tt_ind_solve_power> and :func:sm.stats.zt_ind_solve_power <zt_ind_solve_power> to solve for any one of the parameters of the power equations. See this blog post <http://jpktd.blogspot.fr/2013/03/statistical-power-in-statsmodels.html>_ for a more in-depth description of the additions.
IPython notebook examples: Many of our examples have been converted or added as IPython notebooks now. They are available here <https://www.statsmodels.org/devel/examples/index.html#notebook-examples>_.
Improved marginal effects for discrete choice models: Expanded options for obtaining marginal effects after the estimation of nonlinear discrete choice models are available. See :py:meth:get_margeff <statsmodels.discrete.discrete_model.DiscreteResuls.get_margeff>.
OLS influence outlier measures: After the estimation of a model with OLS, the common set of influence and outlier measures and a outlier test are now available attached as methods get_influnce and outlier_test to the Results instance. See :py:class:OLSInfluence <statsmodels.stats.outliers_influence.OLSInfluence> and :func:outlier_test <statsmodels.stats.outliers_influence.outlier_test>.
New datasets: New :ref:datasets <datasets> are available for examples.
Access to R datasets: We now have access to many of the same datasets available to R users through the Rdatasets project <https://vincentarelbundock.github.io/Rdatasets/>_. You can access these using the :func:sm.datasets.get_rdataset <statsmodels.datasets.get_rdataset> function. This function also includes caching of these datasets.
Improved numerical differentiation tools: Numerical differentiation routines have been greatly improved and expanded to cover all the routines discussed in::
Ridout, M.S. (2009) Statistical applications of the complex-step method of numerical differentiation. The American Statistician, 63, 66-74
See the :ref:sm.tools.numdiff <numdiff> module.
Consistent constant handling across models: Result statistics no longer rely on the assumption that a constant is present in the model.
Missing value handling across models: Users can now control what models do in the presence of missing values via the missing keyword available in the instantiation of every model. The options are 'none', 'drop', and 'raise'. The default is 'none', which does no missing value checks. To drop missing values use 'drop'. And 'raise' will raise an error in the presence of any missing data.
.. currentmodule:: statsmodels.iolib
.dta files. See :class:sm.iolib.StataWriter <foreign.StataWriter>... currentmodule:: statsmodels.tsa.arima.model
ARIMA modeling: statsmodels now has support for fitting Autoregressive Integrated Moving Average (ARIMA) models. See :class:ARIMA and :class:ARIMAResults for more information.
Support for dynamic prediction in AR(I)MA models: It is now possible to obtain dynamic in-sample forecast values in :class:ARMA and :class:ARIMA models.
Improved Pandas integration: statsmodels now supports all frequencies available in pandas for time-series modeling. These are used for intelligent dates handling for prediction. These features are available, if you pass a pandas Series or DataFrame with a DatetimeIndex to a time-series model.
.. currentmodule:: statsmodels
interrater), statistics and hypothesis tests for proportions (See :ref:proportion stats <proportion_stats>), Tukey HSD (with plot) was added as an enhancement to the multiple comparison tests (:class:sm.stats.multicomp.MultiComparison <sandbox.stats.multicomp.MultiComparison>, :func:sm.stats.multicomp.pairwise_tukeyhsd <stats.multicomp.pairwise_tukeyhsd>). Weighted statistics and t tests were enhanced with new options. Tests of equivalence for one sample and two independent or paired samples were added based on t tests and z tests (See :ref:tost).Post-estimation statistics for weighted least squares that depended on the centered total sum of squares were not correct. These are now correct and tested. See :issue:501.
Regression through the origin models now correctly use uncentered total sum of squares in post-estimation statistics. This affected the :math:R^2 value in linear models without a constant. See :issue:27.
Cython code is now non-optional. You will need a C compiler to build from source. If building from github and not a source release, you will also need Cython installed. See the :ref:installation documentation <install>.
The q_matrix keyword to t_test and f_test for linear models is deprecated. You can now specify linear hypotheses using formulas.
.. currentmodule:: statsmodels.tsa
The conf_int keyword to :func:sm.tsa.acf <stattools.acf> is deprecated.
The names argument is deprecated in :class:sm.tsa.VAR <vector_ar.var_model.VAR> and sm.tsa.SVAR <vector_ar.svar_model.SVAR>. This is now automatically detected and handled.
.. currentmodule:: statsmodels.tsa
order keyword to :py:meth:sm.tsa.ARMA.fit <ARMA.fit> is deprecated. It is now passed in during model instantiation... currentmodule:: statsmodels.distributions
The empirical distribution function (:class:sm.distributions.ECDF <ECDF>) and supporting functions have been moved to statsmodels.distributions. Their old paths have been deprecated.
The margeff method of the discrete choice models has been deprecated. Use get_margeff instead. See above. Also, the vague resid attribute of the discrete choice models has been deprecated in favor of the more descriptive resid_dev to indicate that they are deviance residuals.
.. currentmodule:: statsmodels.nonparametric.kde
KDE has been deprecated and renamed to :class:KDEUnivariate to distinguish it from the new KDEMultivariate. See above.The previous version (statsmodels 0.4.3) was released on July 2, 2012. Since then we have closed a total of 380 issues, 172 pull requests and 208 regular issues. The :ref:detailed list<issues_list_05> can be viewed.
This release is a result of the work of the following 38 authors who contributed total of 2032 commits. If for any reason, we have failed to list your name in the below, please contact us:
.. note::
Obtained by running git log v0.4.3..HEAD --format='* %aN <%aE>' | sed 's/@/\-at\-/' | sed 's/<>//' | sort -u.
.. _issues_list_05:
GitHub stats for release 0.5.0 (07/02/2012/ - 08/14/2013/).
We closed a total of 380 issues, 172 pull requests and 208 regular issues. This is the full list (generated with the script :file:tools/github_stats.py):
This list is automatically generated, and may be incomplete:
Pull Requests (172):
1015: DOC: Bump version. Remove done tasks.1010: DOC/RLS: Update release notes workflow. Help Needed!1014: DOC: nbgenerate does not like the comment at end of line.1012: DOC: Add link to notebook and crosslink ref. Closes #924.997: misc, tests, diagnostic1009: MAINT: Add .mailmap file.817: Add 3 new unit tests for arima_process1001: BUG include_package_data for install closes #9071005: GITHUB: Contributing guidelines1007: Cleanup docs for release1003: BUG: Workaround for bug in sphinx 1.1.3. See #1002.1004: DOC: Update maintainer notes with branching instructions.1000: BUG: Support pandas 0.8.0.996: BUG: Handle combo of pandas 0.8.0 and dateutils 1.5.0995: ENH: Print dateutil version.994: ENH: Fail gracefully for version not found.993: More conservative error catching in TimeSeriesModel992: Misc fixes 12: adjustments to unit test985: MAINT: Print versions script.986: ENH: Prefer to_offset to get_offset. Closes #964.984: COMPAT: Pandas 0.8.1 compatibility. Closes #983.982: Misc fixes 11978: TST: generic mle pareto disable bsejac tests with estimated loc977: BUG python 3.3 fix for numpy str TypeError, see #633975: Misc fixes 10 numdiff970: BUG: array too long, raises exception with newer numpy closes #967965: Vincent summary2 rebased933: Update and improve GenericlikelihoodModel and miscmodels950: BUG/REF mcnemar fix exact pvalue, allow table as input951: Pylint emplike formula genmod956: Fix a docstring in KDEMultivariateConditional.949: BUG fix lowess sort when nans closes #946932: ENH: support basinhopping solver in LikelihoodModel.fit()927: DOC: clearer minimal example919: OLS summary crash918: Fixes10 emplike lowess909: Bugs in GLM pvalues, more tests, pylint906: ENH: No fmax with Windows SDK so define inline.905: MAINT more fixes898: Misc fixes 7896: Quantreg rebase2895: Fixes issue #832893: ENH: Remove unneeded restriction on low. Closes #867.894: MAINT: Remove broken function. Keep deprecation. Closes #781.856: Carljv improved lowess rebased2884: Pyflakes cleanup887: BUG: Fix kde caching883: Fixed pyflakes issue in discrete module882: Update predstd.py871: Update of sandbox doc631: WIP: Correlation positive semi definite857: BLD: apt get dependencies from Neurodebian, whitespace cleanup855: AnaMP issue 783 mixture rvs tests rebased854: Enrico multinear rebased849: Tyler tukeyhsd rebased848: BLD TravisCI use python-dateutil package784: Misc07 cleanup multipletesting and proportions841: ENH: Add load function to main API. Closes #840.820: Ensure that tuples are not considered as data, not as data containers822: DOC: Update for Cython changes.765: Fix build issues800: Automatically generate output from notebooks802: BUG: Use two- not one-sided t-test in t_test. Closes #740.806: ENH: Import formula.api in statsmodels.api namespace.803: ENH: Fix arima error message for bad start_params801: DOC: Fix ANOVA section titles795: Negative Binomial Rebased787: Origintests794: ENH: Allow pandas-in/pandas-out in tsa.filters791: Github stats for release notes779: added np.asarray call to durbin_watson in stattools772: Anova docs776: BUG: Fix dates_from_range with length. Closes #775.774: BUG: Attach prediction start date in AR. Closes #773.767: MAINT: Remove use of deprecated from examples and docs.762: ENH: Add new residuals to wrapper754: Fix arima predict760: ENH: Adjust for k_trend in information criteria. Closes #324.761: ENH: Fixes and tests sign_test. Closes #642.759: Fix 236758: DOC: Update VAR docs. Closes #537.752: Discrete cleanup750: VAR with 1d array748: Remove reference to new_t_test and new_f_test.739: DOC: Remove outdated note in docstring732: BLD: Check for patsy dependency at build time + docs731: Handle wrapped730: Fix opt fulloutput729: Get rid of warnings in docs build698: update url for hsb2 dataset727: DOC: Fix indent and add missing params to linear models. Closes #709.726: CLN: Remove unused method. Closes #694725: BUG: Should call anova_single. Closes #702.723: Rootfinding for Power722: Handle pandas.Series with names in make_lags714: Fix 712668: Allow for any pandas frequency to be used in TimeSeriesModel.711: Misc06 - bug fixes708: BUG: Fix one regressor case for conf_int. Closes #706.700: Bugs rebased680: BUG: Swap arguments in fftconvolve for scipy >= 0.12.0640: Misc fixes 05663: a typo in runs.py doc string for mcnemar test652: WIP: fixing pyflakes / pep8, trying to improve readability619: DOC: intro to formulas648: BF: Make RLM stick to Huber's description649: Bug Fix637: Pyflakes cleanup634: VAR DOC typo623: Slowtests621: MAINT: in setup.py, only catch ImportError for pandas.590: Cleanup test output591: Interrater agreement and reliability measures618: Docs fix the main warnings and errors during sphinx build610: nonparametric examples and some fixes578: Fix 577575: MNT: Remove deprecated scikits namespace499: WIP: Handle constant567: Remove deprecated571: Dataset docs561: Grab rdatasets570: DOC: Fixed links to Rdatasets524: DOC: Clean up discrete model documentation.506: ENH: Re-use effects if model fit with QR556: WIP: L1 doc fix564: TST: Use native integer to avoid issues in dtype asserts543: Travis CI using M.Brett nipy hack558: Plot cleanup541: Replace pandas DataMatrix with DataFrame534: Stata test fixes532: Compat 323531: DOC: Add ECDF to distributions docs526: ENH: Add class to write Stata binary dta files521: DOC: Add abline plot to docs518: Small fixes: interaction_plot508: ENH: Avoid taking cholesky decomposition of diagonal matrix509: DOC: Add ARIMA to docs510: DOC: realdpi is disposable personal income. Closes #394.507: ENH: Protect numdifftools import. Closes #45504: Fix weights498: DOC: Add patys requirement to install docs491: Make _data a public attribute.494: DOC: Fix pandas links492: added intersphinx for pandas422: Handle missing data485: ENH: Improve error message for pandas objects without dates in index428: Remove other data483: Arima predict bug482: TST: Do array-array comparison when using numpy.testing471: Formula rename df -> data473: Vincent docs tweak rebased468: Docs 050462: El aft rebased461: TST: numpy 1.5.1 compatibility460: Emplike desc reg rebase410: Discrete model marginal effects417: Numdiff cleanup398: Improved plot_corr and plot_corr_grid functions.401: BUG: Finish refactoring margeff for dummy. Closes #399.400: MAINT: remove lowess.py, which was kept in 0.4.x for backwards compatibi...371: BF+TEST: fixes, checks and tests for isestimable351: ENH: Copy diagonal before write for upcoming numpy changes384: REF: Move mixture_rvs out of sandbox.368: ENH: Add polished version of acf/pacf plots with confidence intervals378: Infer freq374: ENH: Add Fair's extramarital affair dataset. From tobit-model branch.358: ENH: Add method to OLSResults for outlier detection369: ENH: allow predict to pass through patsy for transforms352: Formula integration rebased360: REF: Deprecate order in fit and move to ARMA init366: Version fixes359: DOC: Fix sphinx warningsIssues (208):
1036: Series no longer inherits from ndarray1038: DataFrame with integer names not handled in ARIMA1028: Test fail with windows and Anaconda - Low priority676: acorr_breush_godfrey undefined nlags922: lowess returns inconsistent with option425: no bse in robust with norm=TrimmedMean1025: add_constant incorrectly detects constant column533: py3 compatibility pandas.read_csv(urlopen(...))662: doc: install instruction: explicit about removing scikits.statsmodels910: test failure Ubuntu TestARMLEConstant.test_dynamic_predict80: t_model: f_test, t_test do not work432: GenericLikelihoodModel change default for score and hessian454: BUG/ENH: HuberScale instance is not used, allow user defined scale estimator98: check connection or connect summary to variable names in wrappers418: BUG: MNLogit loglikeobs, jac1017: nosetests warnings924: DOCS link in notebooks to notebook for download1011: power ttest endless loop possible907: BLD data_files for stats.libqsturng328: consider moving example scripts into IPython notebooks1002: Docs will not build with Sphinx 1.1.369: Make methods like compare_ftest work with wrappers503: summary_old in RegressionResults991: TST precision of normal_power945: Installing statsmodels from github?964: Prefer to_offset not get_offset in tsa stuff983: bug: pandas 0.8.1 incompatibility899: build_ext inplace does not cythonize923: location of initialization code980: auto lag selection in S_hac_simple968: genericMLE Ubuntu test failure633: python 3.3 compatibility728: test failure for solve_power with fsolve971: numdiff test cases976: VAR Model does not work in 1D972: numdiff: epsilon has no minimum value967: lowes test failure Ubuntu948: nonparametric tests: mcnemar, cochranq unit test963: BUG in runstest_2sample946: Issue with lowess() smoother in statsmodels868: k_vars > nobs917: emplike emplikeAFT stray dimensions264: version comparisons need to be made more robust (may be just use LooseVersion)674: failure in test_foreign, pandas testing828: GLMResults inconsistent distribution in pvalues908: RLM missing test for tvalues, pvalues463: formulas missing in docs256: discrete Nbin has zero test coverage831: test errors running bdist733: Docs: interrater cohens_kappa is missing897: lowess failure - sometimes902: test failure tsa.filters precision too high901: test failure stata_writer_pandas, newer versions of pandas900: ARIMA.new errors on python 3.3832: notebook errors867: Baxter King has unneeded limit on value for low?781: discreteResults margeff method not tests, obsolete870: discrete unit tests duplicates630: problems in regression plots885: Caching behavior for KDEUnivariate icdf869: sm.tsa.ARMA(..., order=(p,q)) gives "init() got an unexpected keyword argument 'order'" error783: statsmodels.distributions.mixture_rvs.py no unit tests824: Multicomparison w/Pandas Series789: presentation of multiple comparison results764: BUG: multipletests incorrect reject for Holm-Sidak766: multipletests - status and tests of 2step FDR procedures763: Bug: multipletests raises exception with empty array840: sm.load should be in the main API namespace830: invalid version number821: Fail gracefully when extensions are not built204: Cython extensions built twice?689: tutorial notebooks740: why does t_test return one-sided p-value804: What goes in statsmodels.formula.api?675: Improve error message for ARMA SVD convergence failure.15: arma singular matrix559: Add Rdatasets to optional dependencies list796: Prediction Standard Errors793: filters are not pandas aware785: Negative R-squared777: OLS residuals returned as Pandas series when endog and exog are Pandas series770: Add ANOVA to docs775: Bug in dates_from_range773: AR model pvalues error with Pandas768: multipletests: numerical problems at threshold355: add draw if interactive to plotting functions625: Exog is not correctly handled in ARIMA predict626: ARIMA summary does not print exogenous variable coefficients657: order (0,1) breaks ARMA forecast736: ARIMA predict problem for ARMA model324: ic in ARResults, aic, bic, hqic, fpe inconsistent definition?642: sign_test check236: AR start_params broken235: tests hang on Windows156: matplotlib deprecated legend ? var plots331: Remove stale tests592: test failures in datetools537: Var Models755: Unable to access AR fit parameters when model is estimated with pandas.DataFrame670: discrete: numerically useless clipping515: MNLogit residuals raise a TypeError225: discrete models only define deviance residuals594: remove skiptest in TestProbitCG681: Dimension Error in discrete_model.py When Running test_dummy_*744: DOC: new_f_test549: Ship released patsy source in statsmodels588: patsy is a hard dependency?716: Tests missing for functions if pandas is used715: statsmodels regression plots not working with pandas datatypes450: BUG: full_output in optimizers Likelihood model709: DOCstrings linear models do not have missing params370: BUG weightstats has wrong cov694: DiscreteMargins duplicate method702: bug, pylint stats.anova423: Handling of constant across models456: BUG: ARMA date handling incompatibility with recent pandas514: NaNs in Multinomial405: Check for existing old version of scikits.statsmodels?586: Segmentation fault with OLS721: Unable to run AR on named time series objects125: caching pinv_wexog broke iterative fit - GLSAR712: TSA bug with frequency inference319: Timeseries Frequencies707: .summary with alpha ignores parsed value673: nonparametric: bug in _kernel_base710: test_power failures706: .conf_int() fails on linear regression without intercept679: Test Baxter King band-pass filter fails with scipy 0.12 beta1552: influence outliers breaks when regressing on constant639: test folders not on python path565: omni_normtest does not propagate the axis argument563: error in doc generation for AR.fit109: TestProbitCG failure on Ubuntu661: from scipy import comb fails on the latest scipy 0.11.0413: DOC: example_discrete.py missing from 0.5 documentation644: FIX: factor plot + examples broken645: STY: pep8 violations in many examples173: doc sphinx warnings601: bspline.py dependency on old scipy.stats.models103: ecdf and step function conventions18: Newey-West sandwich covariance is missing279: cov_nw_panel not tests, example broken150: precision in test_discrete.TestPoissonNewton.test_jac ?480: rescale loglike for optimization627: Travis-CI support for scipy622: mark tests as slow in emplike589: OLS F-statistic error572: statsmodels/tools/data.py Stuck looking for la.py580: test errors in graphics577: PatsyData detection buglet470: remove deprecated features573: lazy imports are (possibly) very slow438: New results instances are not in online documentation542: Regression plots fail when Series objects passed to sm.OLS239: release 0.4.x530: l1 docs issues539: test for statawriter (failure)490: Travis CI on PRs252: doc: distributions.rst refers to sandbox only85: release 0.465: MLE fit of AR model has no tests522: test does not propagate arguments to nose517: missing array conversion or shape in linear model523: test failure with ubuntu decimals too large520: web site documentation, source not updated488: Avoid cholesky decomposition of diagonal matrices in linear regression models394: Definition in macrodata NOTE45: numdifftools dependency501: WLS/GLS post estimation results500: WLS fails if weights is a pandas.Series27: add hasconstant indicator for R-squared and df calculations497: DOC: add patsy?495: ENH: add footer SimpleTable402: model._data -> model.data?477: VAR NaN Bug421: Enhancement: Handle Missing Data489: Expose model._data as model.data315: tsa models assume pandas object indices are dates440: arima predict is broken for steps > q and q != 1458: TST BUG? comparing pandas and array in tests, formula464: from_formula signature245: examples in docs: make nicer466: broken example, pandas57: Unhelpful error from bad exog matrix in model.py271: ARMA.geterrors requires model to be fit350: Writing to array returned np.diag354: example_rst does not copy unchanged files over467: Install issues with Pandas444: ARMA example on stable release website not working377: marginal effects count and discrete adjustments426: "svd" method not supported for OLS.fit()409: Move numdiff out of the sandbox416: Switch to complex-step Hessian for AR(I)MA415: bug in kalman_loglike_complex397: plot_corr axis text labeling not working (with fix)399: discrete errors due to incorrect in-place operation389: VAR test_normality is broken with KeyError388: Add tsaplots to graphics.api as graphics.tsa387: predict date was not getting set with start = None386: p-values not returned from acf385: Allow AR.select_order to work without model being fit383: Move mixture_rvs out of sandbox.248: ARMA breaks with a 1d exog273: When to give order for AR/AR(I)MA363: examples folder -> tutorials folder346: docs in sitepackages353: PACF docs raise a sphinx warning348: python 3.2.3 test failure zip_longest