docs/source/release/version0.6.rst
:orphan:
statsmodels 0.6.1 is a bugfix release. All users are encouraged to upgrade to 0.6.1.
See the :ref:list of fixed issues <issues_list_06> for specific backported fixes.
statsmodels 0.6.0 is another large release. It is the result of the work of 37 authors over the last year and includes over 1500 commits. It contains many new features, improvements, and bug fixes detailed below.
See the :ref:list of fixed issues <issues_list_06> for specific closed issues.
The following major new features appear in this version.
Generalized Estimating Equations (GEE) provide an approach to handling dependent data in a regression analysis. Dependent data arise commonly in practice, such as in a longitudinal study where repeated observations are collected on subjects. GEE can be viewed as an extension of the generalized linear modeling (GLM) framework to the dependent data setting. The familiar GLM families such as the Gaussian, Poisson, and logistic families can be used to accommodate dependent variables with various distributions.
Here is an example of GEE Poisson regression in a data set with four count-type repeated measures per subject, and three explanatory covariates.
.. code-block:: python
import numpy as np import statsmodels.api as sm import statsmodels.formula.api as smf
data = sm.datasets.get_rdataset("epil", "MASS").data
md = smf.gee("y ~ age + trt + base", "subject", data, cov_struct=sm.cov_struct.Independence(), family=sm.families.Poisson()) mdf = md.fit() print(mdf.summary())
The dependence structure in a GEE is treated as a nuisance parameter and is modeled in terms of a "working dependence structure". The statsmodels GEE implementation currently includes five working dependence structures (independent, exchangeable, autoregressive, nested, and a global odds ratio for working with categorical data). Since the GEE estimates are not maximum likelihood estimates, alternative approaches to some common inference procedures have been developed. The statsmodels GEE implementation currently provides standard errors, Wald tests, score tests for arbitrary parameter contrasts, and estimates and tests for marginal effects. Several forms of standard errors are provided, including robust standard errors that are approximately correct even if the working dependence structure is misspecified.
Adding functionality to look at seasonality in plots. Two new functions are :func:sm.graphics.tsa.month_plot and :func:sm.graphics.tsa.quarter_plot. Another function :func:sm.graphics.tsa.seasonal_plot is available for power users.
.. code-block:: python
import statsmodels.api as sm
import pandas as pd
dta = sm.datasets.elnino.load_pandas().data
dta['YEAR'] = dta.YEAR.astype(int).astype(str)
dta = dta.set_index('YEAR').T.unstack()
date_str = [f"{yr}-{mo}-01" for yr, mo in dta.index.values]
dta.index = pd.DatetimeIndex(pd.to_datetime(date_str,format="%Y-%b-%d"), freq='MS')
fig = sm.tsa.graphics.month_plot(dta)
.. currentmodule:: statsmodels.tsa
We added a naive seasonal decomposition tool in the same vein as R's decompose. This function can be found as :func:sm.tsa.seasonal_decompose <tsa.seasonal.seasonal_decompose>.
.. plot:: :include-source:
import statsmodels.api as sm
dta = sm.datasets.co2.load_pandas().data
# deal with missing values. see issue
co2 = dta.co2.interpolate()
res = sm.tsa.seasonal_decompose(co2)
res.plot()
Addition of Linear Mixed Effects Models (MixedLM)
Linear Mixed Effects models are used for regression analyses involving dependent data. Such data arise when working with longitudinal and other study designs in which multiple observations are made on each subject. Two specific mixed effects models are "random intercepts models", where all responses in a single group are additively shifted by a value that is specific to the group, and "random slopes models", where the values follow a mean trajectory that is linear in observed covariates, with both the slopes and intercept being specific to the group. The statsmodels MixedLM implementation allows arbitrary random effects design matrices to be specified for the groups, so these and other types of random effects models can all be fit.
Here is an example of fitting a random intercepts model to data from a longitudinal study:
.. code-block:: python
import statsmodels.api as sm
import statsmodels.formula.api as smf
data = sm.datasets.get_rdataset('dietox', 'geepack', cache=True).data
md = smf.mixedlm("Weight ~ Time", data, groups=data["Pig"])
mdf = md.fit()
print(mdf.summary())
The statsmodels LME framework currently supports post-estimation inference via Wald tests and confidence intervals on the coefficients, profile likelihood analysis, likelihood ratio testing, and AIC. Some limitations of the current implementation are that it does not support structure more complex on the residual errors (they are always homoscedastic), and it does not support crossed random effects. We hope to implement these features for the next release.
It is now possible to call out to X-12-ARIMA or X-13ARIMA-SEATS from statsmodels. These libraries must be installed separately.
.. code-block:: python
import statsmodels.api as sm
dta = sm.datasets.co2.load_pandas().data
co2 = dta.co2.interpolate()
co2 = co2.resample('M').last()
res = sm.tsa.x13_arima_select_order(dta.co2)
print(res.order, res.sorder)
results = sm.tsa.x13_arima_analysis(co2)
fig = results.plot()
fig.set_size_inches(12, 5)
fig.tight_layout()
The Kalman filter Cython code underlying AR(I)MA estimation has been substantially optimized. You can expect speed-ups of one to two orders of magnitude.
Added :func:sm.tsa.arma_order_select_ic. A convenience function to quickly get the information criteria for use in tentative order selection of ARMA processes.
Plotting functions for timeseries is now imported under the sm.tsa.graphics namespace in addition to sm.graphics.tsa.
New distributions.ExpandedNormal class implements the Edgeworth expansion for weakly non-normal distributions.
New datasets: Added new :ref:datasets <datasets> for examples. sm.datasets.co2 is a univariate time-series dataset of weekly co2 readings. It exhibits a trend and seasonality and has missing values.
Added robust skewness and kurtosis estimators in :func:sm.stats.stattools.robust_skewness and :func:sm.stats.stattools.robust_kurtosis, respectively. An alternative robust measure of skewness has been added in :func:sm.stats.stattools.medcouple.
New functions added to correlation tools: corr_nearest_factor
finds the closest factor-structured correlation matrix to a given
square matrix in the Frobenius norm; corr_thresholded efficiently
constructs a hard-thresholded correlation matrix using sparse matrix
operations.
New dot_plot in graphics: A dotplot is a way to visualize a small dataset
in a way that immediately conveys the identity of every point in the plot.
Dotplots are commonly seen in meta-analyses, where they are known
as "forest plots", but can be used in many other settings as well.
Most tables that appear in research papers can be represented
graphically as a dotplot.
statsmodels has added custom warnings to statsmodels.tools.sm_exceptions. By default all of these warnings will be raised whenever appropriate. Use warnings.simplefilter to turn them off, if desired.
Allow control over the namespace used to evaluate formulas with patsy via the eval_env keyword argument. See the :ref:patsy-namespaces documentation for more information.
805, :issue:1877.2013.d = 1. :issue:1562... currentmodule:: statsmodels.tsa
statsmodels.tsa.filters.arfilter has been removed. This did not compute a recursive AR filter but was instead a convolution filter. Two new functions have been added with clearer names :func:sm.tsa.filters.recursive_filter <tsa.filters.filtertools.recursive_filter> and :func:sm.tsa.filters.convolution_filter <tsa.filters.filtertools.convolution_filter>.The previous version (0.5.0) was released August 14, 2014. Since then we have closed a total of 528 issues, 276 pull requests, and 252 regular issues. Refer to the :ref:detailed list<issues_list_06> for more information.
This release is a result of the work of the following 37 authors who contributed a total of 1531 commits. If for any reason we have failed to list your name in the below, please contact us:
A blurb about the number of changes and the contributors list.
.. note::
Obtained by running git log v0.5.0..HEAD --format='* %aN <%aE>' | sed 's/@/\-at\-/' | sed 's/<>//' | sort -u.
.. _issues_list_06:
GitHub stats for 2013/08/14 - 2014/10/15 (tag: v0.5.0)
We closed a total of 528 issues, 276 pull requests and 252 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 (276):
2044: ENH: Allow unit interval for binary models. Closes #2040.1426: ENH: Import arima_process stuff into tsa.api2042: Fix two minor typos in contrast.py2034: ENH: Handle missing for extra data with formulas2035: MAINT: Remove deprecated code for 0.61325: ENH: add the Edgeworth expansion based on the normal distribution2032: DOC: What it is what it is.2031: ENH: Expose patsy eval_env to users.2028: ENH: Fix numerical issues in links and families.2029: DOC: Fix versions to match other docs.1647: ENH: Warn on non-convergence.2014: BUG: Fix forecasting for ARIMA with d == 22013: ENH: Better error message on object dtype2012: BUG: 2d 1 columns -> 1d. Closes #322.2009: DOC: Update after refactor. Use code block.2008: ENH: Add wrapper for MixedLM1954: ENH: PHReg formula improvements2007: BLD: Fix build issues2006: BLD: Do not generate cython on clean. Closes #1852.2000: BLD: Let pip/setuptools handle dependencies that are not installed at all.1999: Gee offset exposure 1994 rebased1998: BUG/ENH Lasso emptymodel rebased1989: BUG/ENH: WLS generic robust cov_type did not use whitened,1587: ENH: Wrap X12/X13-ARIMA AUTOMDL. Closes #442.1563: ENH: Add plot_predict method to ARIMA models.1995: BUG: Fix issue #19931981: ENH: Add api for covstruct. Clear init. Closes #1917.1996: DEV: Ignore .venv file.1982: REF: Rename jac -> score_obs. Closes #1785.1987: BUG tsa pacf, base bootstrap1986: Bug multicomp 1927 rebased1984: Docs add gee.rst1985: Bug uncentered latex table 1929 rebased1983: BUG: Fix compat asunicode1574: DOC: Fix math.1980: DOC: Documentation fixes1974: REF/Doc beanplot change default color, add notebook1978: ENH: Check input to binary models1979: BUG: Typo1976: ENH: Add repr_html to SimpleTable1977: BUG: Fix import refactor victim.1975: BUG: Yule walker cast to float1973: REF: Move and expose webuse1972: TST: Add testing against NumPy 1.9 and matplotlib 1.41939: ENH: Binstar build files1952: REF/DOC: Misc1940: REF: refactor and speedup of mixed LME1937: ENH: Quick access to online documentation1942: DOC: Rename Change README type to rst1938: ENH: Enable Python 3.4 testing1924: Bug gee cov type 1906 rebased1870: robust covariance, cov_type in fit1859: BUG: Do not use negative indexing with k_ar == 0. Closes #1858.1914: BUG: LikelihoodModelResults.pvalues use df_resid_inference1899: TST: fix assert_equal for pandas index1895: Bug multicomp pandas1894: BUG fix more ix indexing cases for pandas compat1889: BUG: fix ytick positions closes #15611887: Bug pandas compat asserts1888: TST test_corrpsd Test_Factor: add noise to data1886: BUG pandas 0.15 compatibility in grouputils labels1885: TST: corr_nearest_factor, more informative tests1884: Fix: Add compat code for pd.Categorical in pandas>=0.151883: BUG: add _ctor_param to TransfGen distributions1872: TST: fix _infer_freq for pandas .14+ compat1867: Ref covtype fit1865: Disable tst distribution 18641856: _spg_optim returns history of objective function values1854: BLD: Do not hard-code path for building notebooks. Closes #12491851: MAINT: Cor nearest factor tests1847: Newton regularize1623: BUG Negbin fit regularized1797: BUG/ENH: fix and improve constant detection1770: TST: anova with -1 noconstant, add tests1837: Allow group variable to be passed as variable name when using formula1839: BUG: GEE score1830: BUG/ENH Use t1832: TST error with scipy 0.14 location distribution class1827: fit_regularized for linear models rebase 16741825: Phreg 1312 rebased1826: Lme api docs1824: Lme profile 1695 rebased1823: Gee cat subclass 1694 rebase1781: ENH: Glm add score_obs1821: Glm maint #1734 rebased1820: BUG: revert change to conf_int in PR #18191819: Docwork1772: REF: cov_params allow case of only cov_params_default is defined1771: REF numpy >1.9 compatibility, indexing into empty slice closes #17541769: Fix ttest 1d1766: TST: TestProbitCG increase bound for fcalls closes #16901709: BLD: Made build extensions more flexible1714: WIP: fit_constrained1706: REF: Use fixed params in test. Closes #910.1701: BUG: Fix faulty logic. Do not raise when missing='raise' and no missing data.1699: TST/ENH StandardizeTransform, reparameterize TestProbitCG1697: Fix for statsmodels/statsmodels#16891692: OSL Example: redundant cell in example removed1688: Kshedden mixed rebased of #13981629: Pull request to fix bandwidth bug in issue 5971666: Include pyx in sdist but do not install1683: TST: GLM shorten random seed closes #16821681: Dotplot kshedden rebased of 12941679: BUG: Fix problems with predict handling offset and exposure1677: Update docstring of RegressionModel.predict()1635: Allow offset and exposure to be used together with log link; raise except...1676: Tests for SVAR1671: ENH: avoid hard-listed bandwidths -- use present dictionary (+typos fixed)1643: Allow matrix structure in covariance matrices to be exploited1657: BUG: Fix refactor victim.1630: DOC: typo, "intercept"1619: MAINT: Dataset docs cleanup and automatic build of docs1612: BUG/ENH Fix negbin exposure #16111610: BUG/ENH fix llnull, extra kwds to recreate model1582: BUG: wls_prediction_std fix weight handling, see 9871613: BUG: Fix proportions allpairs #14931607: TST: adjust precision, CI Debian, Ubuntu testing1603: ENH: Allow start_params in GLM1600: CLN: Regression plots fixes1592: DOC: Additions and fixes1520: CLN: Refactored so that there is no longer a need for 2to31585: Cor nearest 1384 rebased1553: Gee maint 1528 rebased1583: BUG: For ARMA(0,0) ensure 1d bse and fix summary.1580: DOC: Fix links. [skip ci]1572: DOC: Fix link title [skip ci]1566: BLD: Fix copy paste path error for >= 3.3 Windows builds1524: ENH: Optimize Cython code. Use scipy blas function pointers.1560: ENH: Allow ARMA(0,0) in order selection1559: MAINT: Recover lost commits from vbench PR1554: Silenced test output introduced in medcouple1234: ENH: Robust skewness, kurtosis and medcouple measures1484: ENH: Add naive seasonal decomposition function1551: COMPAT: Fix failing test on Python 2.61472: ENH: using human-readable group names instead of integer ids in MultiComparison1437: ENH: accept non-int definitions of cluster groups1550: Fix test gmm poisson1549: TST: Fix locally failing tests.1121: WIP: Refactor optimization code.1547: COMPAT: Correct bit_length for 2.61545: MAINT: Fix missed usage of deprecated tools.rank1196: REF: ensure O(N log N) when using fft for acf1154: DOC: Add links for build machines.1546: DOC: Fix link to wrong notebook1383: MAINT: Deprecate rank in favor of np.linalg.matrix_rank1432: COMPAT: Add NumpyVersion from scipy1438: ENH: Option to avoid "center" environment.1544: BUG: Travis miniconda1510: CLN: Improve warnings to avoid generic warnings messages1543: TST: Suppress RuntimeWarning for L-BFGS-B1507: CLN: Silence test output1540: BUG: Correct derivative for exponential transform.1536: BUG: Restores coveralls for a single build1535: BUG: Fixes for 2.6 test failures, replacing astype(str) with apply(str)1523: Travis miniconda1533: DOC: Fix link to code on github1531: DOC: Fix stale links with linkcheck1530: DOC: Fix link1527: DOCS: Update docs add FAQ page1525: DOC: Update with Python 3.4 build notes1518: DOC: Ask for release notes and example.1516: DOC: Update examples contributing docs for current practice.1517: DOC: Be clear about data attribute of Datasets1515: DOC: Fix broken link1514: DOC: Fix formula import convention.1506: BUG: Format and decode errors in Python 2.61505: TST: Test co2 load_data for Python 3.1504: BLD: New R versions require NAMESPACE file. Closes #1497.1483: ENH: Some utility functions for working with dates1482: REF: Prefer filters.api to init1481: ENH: Add weekly co2 dataset1474: DOC: Add plots for standard filter methods.1471: DOC: Fix import1470: DOC/BLD: Log code exceptions from nbgenerate1469: DOC: Fix bad links1468: MAINT: CSS fixes1463: DOC: Remove defunct argument. Change default kw. Closes #1462.1452: STY: import pandas as pd1458: BUG/BLD: exclude sandbox in relative path, not absolute1447: DOC: Only build and upload docs if we need to.1445: DOCS: Example landing page1436: DOC: Fix auto doc builds.1431: DOC: Add default for getenv. Fix paths. Add print_info1429: MAINT: Use ip_directive shipped with IPython1427: TST: Make tests fit quietly1424: ENH: Consistent results for transform_slices1421: ENH: Add grouping utilities code1419: Gee 1314 rebased1414: TST temporarily rename tests probplot other to skip them1403: Bug norm expan shapes1417: REF: Let subclasses keep kwds attached to data.1416: ENH: Make handle_data overwritable by subclasses.1410: ENH: Handle missing is none1402: REF: Expose missing data handling as classmethod1387: MAINT: Fix failing tests1406: MAINT: Tools improvements1404: Tst fix genmod link tests1396: REF: Multipletests reduce memory usage1380: DOC :Update vector_ar.rst1381: BLD: Do not check dependencies on egg_info for pip. Closes #1267.1302: BUG: Fix typo.1375: STY: Remove unused imports and comment out unused libraries in setup.py1143: DOC: Update backport notes for new workflow.1374: ENH: Import tsaplots into tsa namespace. Closes #1359.1369: STY: Pep-8 cleanup1370: ENH: Support ARMA(0,0) models.1368: STY: Pep 8 cleanup1367: ENH: Make sure mle returns attach to results.1365: STY: Import and pep 8 cleanup1364: ENH: Get rid of hard-coded lbfgs. Closes #988.1363: BUG: Fix typo.1361: ENH: Attach mlefit to results not model.1360: ENH: Import adfuller into tsa namespace1346: STY: PEP-8 Cleanup1344: BUG: Use missing keyword given to ARMA.1340: ENH: Protect against ARMA convergence failures.1334: ENH: ARMA order select convenience function1339: Fix typos1336: REF: Get rid of plain assert.1333: STY: all should be after imports.1332: ENH: Add Bunch object to tools.1331: ENH: Always use unicode.1329: BUG: Decode metadata to utf-8. Closes #1326.1330: DOC: Fix typo. Closes #1327.1185: Added support for pandas when pandas was installed directly from git trunk1315: MAINT: Change back to path for build box1305: TST: Update hard-coded path.1290: ENH: Add seasonal plotting.1296: BUG/TST: Fix ARMA forecast when start == len(endog). Closes #12951292: DOC: cleanup examples folder and webpage1286: Make sure PeriodIndex passes through tsa. Closes #1285.1271: Silverman enhancement - Issue #12431264: Doc work GEE, GMM, sphinx warnings1179: REF/TST: ProbPlot now uses resettable_cache and added some kwargs to plotting fxns1225: Sandwich mle1258: Gmm new rebased1255: ENH add GEE to genmod1254: REF: Results.predict convert to array and adjust shape1192: TST: enable tests for llf after change to WLS.loglike see #11701253: Wls llf fix1233: sandbox kernels bugs uniform kernel and confint1240: Kde weights 1103 8231228: Add default value tags to adfuller() docs1198: fix typo1230: BUG: numerical precision in resid_pearson with perfect fit #12291214: Compare lr test rebased1200: BLD: do not install *.pyx *.c MANIFEST.in1202: MAINT: Sort backports to make applying easier.1157: Tst precision1161: add a fitting interface for simultaneous log likelihood and score, for lbfgs, tested with MNLogit1160: DOC: update scipy version from 0.7 to 0.9.01147: ENH: add lbfgs for fitting1156: ENH: Raise on 0,0 order models in AR(I)MA. Closes #11231149: BUG: Fix small data issues for ARIMA.1092: Fixed duplicate svd in RegressionModel1139: TST: Silence tests1135: Misc style1088: ENH: add predict_prob to poisson1125: REF/BUG: Some GLM cleanup. Used trimmed results in NegativeBinomial variance.1124: BUG: Fix ARIMA prediction when fit without a trend.1118: DOC: Update gettingstarted.rst1117: Update ex_arma2.py1107: REF: Deprecate stand_mad. Add center keyword to mad. Closes #658.1089: ENH: exp(poisson.logpmf()) for poisson better behaved.1077: BUG: Allow 1d exog in ARMAX forecasting.1075: BLD: Fix build issue on some versions of easy_install.1071: Update setup.py to fix broken install on OSX1052: DOC: Updating contributing docs1136: RLS: Add IPython tools for easier backporting of issues.1091: DOC: minor git typo1082: coveralls support1072: notebook examples title cell1056: Example: reg diagnostics1057: COMPAT: Fix py3 caching for get_rdatasets.1045: DOC/BLD: Update from nbconvert to IPython 1.0.1026: DOC/BLD: Add LD_LIBRARY_PATH to env for docs build.Issues (252):
2040: enh: fractional Logit, Probit1220: missing in extra data (example sandwiches, robust covariances)1877: error with GEE on missing data.805: nan with categorical in formula2036: test in links require exact class so Logit cannot work in place of logit2010: Go over deprecations again for 0.6.1303: patsy library not automatically installed2024: genmod Links numerical improvements2025: GEE requires exact import for cov_struct2017: Matplotlib warning about too many figures724: check warnings1562: ARIMA forecasts are hard-coded for d=1880: DataFrame with bool type not cast correctly.1992: MixedLM style322: acf / pacf do not work on pandas objects1317: AssertionError: attr is not equal [dtype]: dtype('object') != dtype('datetime64[ns]')1875: dtype bug object arrays (raises in clustered standard errors code)1842: dtype object, glm.fit() gives AttributeError: sqrt1300: Doc errors, missing1164: RLM cov_params, t_test, f_test do not use bcov_scaled1019: 0.6.0 Roadmap554: Prediction Standard Errors333: ENH tools: squeeze in R export file1990: MixedLM does not have a wrapper1897: Consider depending on setuptools in setup.py2003: pip install now fails silently1852: do not cythonize when cleaning up1991: GEE formula interface does not take offset/exposure442: Wrap x-12 arima1993: MixedLM bug1917: API: GEE access to genmod.covariance_structure through api1785: REF: rename jac -> score_obs1969: pacf has incorrect standard errors for lag 01434: A small bug in GenericLikelihoodModelResults.bootstrap()1408: BUG test failure with tsa_plots1337: DOC: HCCM are now available for WLS546: influence and outlier documentation1532: DOC: Related page is out of date1386: Add minimum matplotlib to docs1068: DOC: keeping documentation of old versions on sourceforge329: link to examples and datasets from module pages1804: PDF documentation for statsmodels202: Extend robust standard errors for WLS/GLS1519: Link to user-contributed examples in docs1053: inconvenient: logit when endog is (1,2) instead of (0,1)1555: SimpleTable: add repr html for ipython notebook1366: Change default start_params to .1 in ARMA1869: yule_walker (from statsmodels.regression) raises exception when given an integer array1651: statsmodels.tsa.ar_model.ARResults.predict1738: GLM robust sandwich covariance matrices1779: Some directories under statsmodels dont have _init.py1242: No support for (0, 1, 0) ARIMA Models1571: expose webuse, use cache1860: ENH/BUG/DOC: Bean plot should allow for separate widths of bean and violins.1831: TestRegressionNM.test_ci_beta2 i386 AssertionError1079: bugfix release 0.5.11338: Raise Warning for HCCM use in WLS/GLS1430: scipy min version / issue276: memoize, last argument wins, how to attach sandwich to Results?1943: REF/ENH: LikelihoodModel.fit optimization, make hessian optional1957: BUG: Re-create OLS model using _init_keys1905: Docs: online docs are missing GEE1898: add python 3.4 to continuous integration testing1684: BUG: GLM NegativeBinomial: llf ignores offset and exposure1256: REF: GEE handling of default covariance matrices1760: Changing covariance_type on results1906: BUG: GEE default covariance is not used1931: BUG: GEE subclasses NominalGEE do not work with pandas exog1904: GEE Results does not have a Wrapper1918: GEE: required attributes missing, df_resid1919: BUG GEE.predict uses link instead of link.inverse1858: BUG: arimax forecast should special case k_ar == 01903: BUG: pvalues for cluster robust, with use_t do not use df_resid_inference1243: kde silverman bandwidth for non-gaussian kernels1866: Pip dependencies1850: TST test_corr_nearest_factor fails on Ubuntu292: python 3 examples1868: ImportError: No module named compat [ from statsmodels.compat import lmap ]1890: BUG tukeyhsd nan in group labels1891: TST test_gmm outdated pandas, compat1561: BUG plot for tukeyhsd, MultipleComparison1864: test failure sandbox distribution transformation with scipy 0.14.0576: Add contributing guidelines1873: GenericLikelihoodModel is not picklable1822: TST failure on Ubuntu pandas 0.14.0 , problems with frequency1249: Source directory problem for notebook examples1855: anova_lm throws error on models created from api.ols but not formula.api.ols1853: a large number of hardcoded paths1792: R² adjusted strange after including interaction term1794: REF: has_constant, k_constant, include implicit constant detection in base1454: NegativeBinomial missing fit_regularized method1615: REF DRYing fit methods1453: Discrete NegativeBinomialModel regularized_fit ValueError: matrices are not aligned1836: BUG Got an TypeError trying to import statsmodels.api1829: BUG: GLM summary show "t" use_t=True for summary1828: BUG summary2 does not propagate/use use_t1812: BUG/ REF conf_int and use_t1835: Problems with installation using easy_install1801: BUG 'f_gen' missing in scipy 0.14.01803: Error revealed by numpy 1.9.0r11834: stackloss1728: GLM.fit maxiter=0 incorrect1795: singular design with offset ?1730: ENH/Bug cov_params, generalize, avoid ValueError1754: BUG/REF: assignment to slices in numpy >= 1.9 (emplike)1409: GEE test errors on Debian Wheezy1521: ubuntu failures: tsa_plot and grouputils1415: test failure test_arima.test_small_data1213: df_diff in anova_lm1323: Contrast Results after t_test summary broken for 1 parameter109: TestProbitCG failure on Ubuntu1690: TestProbitCG: 8 failing tests (Python 3.4 / Ubuntu 12.04)1763: Johansen method does not give correct index values1761: doc build failures: ipython version ? ipython directive1762: Unable to build1745: UnicodeDecodeError raised by get_rdataset("Guerry", "HistData")611: test failure foreign with pandas 0.7.31700: faulty logic in missing handling1648: ProbitCG failures1689: test_arima.test_small_data: SVD fails to converge (Python 3.4 / Ubuntu 12.04)597: BUG: nonparametric: kernel, efficient=True changes bw even if given1606: BUILD from sdist broken if cython available1246: test failure test_anova.TestAnova2.test_results50: t_test, f_test, model.py for normal instead of t-distribution1655: newey-west different than R?1682: TST test failure on Ubuntu, random.seed1614: docstring for regression.linear_model.RegressionModel.predict() does not match implementation1318: GEE and GLM scale parameter519: L1 fit_regularized cleanup, comments651: add structure to example page1067: Kalman Filter convergence. How close is close enough?1281: Newton convergence failure prints warnings instead of warning1628: Unable to install statsmodels in the same requirements file as numpy, pandas, etc.617: Problem in installing statsmodels in Fedora 17 64-bit935: ll_null in likelihoodmodels discrete704: datasets.sunspot: wrong link in description1222: NegativeBinomial ignores exposure1611: BUG NegativeBinomial ignores exposure and offset1608: BUG: NegativeBinomial, llnul is always default 'nb2'1221: llnull with exposure ?1493: statsmodels.stats.proportion.proportions_chisquare_allpairs has hardcoded value1260: GEE test failure on Debian1261: test failure on Debian443: GLM.fit does not allow start_params1602: Fitting GLM with a pre-assigned starting parameter1601: Fitting GLM with a pre-assigned starting parameter890: regression_plots problems (pylint) and missing test coverage1598: Is "old" string formatting Python 3 compatible?1589: AR vs ARMA order specification1134: Mark knownfails1259: Parameterless models616: python 2.6, python 3 in single codebase1586: Kalman Filter errors with new pyx1565: build_win_bdist_py3.bat are using the wrong compiler843: UnboundLocalError When trying to install OS X713: arima.fit performance367: unable to install on RHEL 5.61548: testtransf error1478: is sm.tsa.filters.arfilter an AR filter?1420: GMM poisson test failures1145: test_multi noise1539: NegativeBinomial strange results with bfgs936: vbench for statsmodels1153: Where are all our testing machines?1500: Use Miniconda for test builds1526: Out of date docs1311: BUG/BLD 3.4 compatibility of cython c files1513: build on osx -python-3.41497: r2nparray needs NAMESPACE file1502: coveralls coverage report for files is broken1501: pandas in/out in predict1494: truncated violin plots1443: Crash from python.exe using linear regression of statsmodels1462: qqplot line kwarg is broken/docstring is wrong1457: BUG/BLD: Failed build if "sandbox" anywhere in statsmodels path1441: wls function: syntax error "unexpected EOF while parsing" occurs when name of dependent variable starts with digits1428: ipython_directive does not work with ipython1385: SimpleTable in Summary (e.g. OLS) is slow for large models1399: UnboundLocalError: local variable 'fittedvalues' referenced before assignment1377: TestAnova2.test_results fails with pandas 0.13.11394: multipletests: reducing memory consumption1267: Packages cannot have both pandas and statsmodels in install_requires1359: move graphics.tsa to tsa.graphics356: docs take up a lot of space988: AR.fit no precision options for fmin_l_bfgs_b990: AR fit with bfgs: large score14: arma with exog1348: reset_index + set_index with drop=False1343: ARMA does not pass missing keyword up to TimeSeriesModel1326: formula example notebook broken1327: typo in docu-code for "Outlier and Influence Diagnostic Measures"1309: Box-Cox transform (some code needed: lambda estimator)1059: sm.tsa.ARMA making ma invertibility1295: Bug in ARIMA forecasting when start is int len(endog) and dates are given1285: tsa models fail on PeriodIndex with pandas1269: KPSS test for stationary processes1268: Feature request: Exponential smoothing1250: DOCs error in var_plots1032: Poisson predict breaks on list347: minimum number of observations - document or check ?1170: WLS log likelihood, aic and bic1187: sm.tsa.acovf fails when both unbiased and fft are True1239: sandbox kernels, problems with inDomain1231: sandbox kernels confint missing alpha1245: kernels cosine differs from Stata823: KDEUnivariate with weights1229: precision problems in degenerate case1219: select_order1206: REF: RegressionResults cov-HCx into cached attributes1152: statsmodels failing tests with pandas1195: pyximport.install() before import api crash1066: gmm.IV2SLS has wrong predict signature1186: OLS when exog is 1d1113: TST: precision too high in test_normality1159: scipy version is still >= 0.7?1108: SyntaxError: unqualified exec is not allowed in function 'test_EvalEnvironment_capture_flag1116: Typo in Example Doc?1123: BUG : arima_model._get_predict_out_of_sample, ignores exogenous of there is no trend ?1155: ARIMA - The computed initial AR coefficients are not stationary979: Win64 binary cannot find Python installation1046: TST: test_arima_small_data_bug on current main1146: ARIMA fit failing for small set of data due to invalid maxlag1081: streamline linear algebra for linear model1138: BUG: pacf_yw does not demean1127: Allow linear link model with Binomial families1122: no data cleaning for statsmodels.genmod.families.varfuncs.NegativeBinomial()658: robust.mad is not being computed correctly or is non-standard definition; it returns the median1076: Some issues with ARMAX forecasting1073: easy_install sandbox violation1115: EasyInstall Problem1106: bug in robust.scale.mad?1102: Installation Problem1084: DataFrame.sort_index does not use ascending when then value is a list with a single element393: marginal effects in discrete choice do not have standard errors defined1078: Use pandas.version.short_version96: deepcopy breaks on ResettableCache1055: datasets.get_rdataset string decode error on python 346: tsa.stattools.acf confint needs checking and tests957: ARMA start estimate with numpy main62: GLSAR incorrect initial condition in whiten1021: from_formula() throws error - problem installing911: noise in stats.power tests472: Update roadmap for 0.5238: release 0.51006: update nbconvert to IPython 1.01038: DataFrame with integer names not handled in ARIMA1036: Series no longer inherits from ndarray1028: 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 column