Statsmodels is a Python package that provides a complement to scipy for statistical computations including descriptive statistics and estimation and inference ...SkiptomaincontentSwitchtomobileversionSearchPyPISearchstatsmodels0.6.1pipinstallstatsmodels==0.6.1CopyPIPinstructionsNewerversionavailable(0.12.2)Released:Dec2,2014StatisticalcomputationsandmodelsforusewithSciPyNavigationProjectdescriptionReleasehistoryDownloadfilesProjectlinksHomepageStatisticsViewstatisticsforthisprojectviaLibraries.io,orbyusingourpublicdatasetonGoogleBigQueryMetaLicense:BSDLicense(BSDLicense)Author:SkipperSeabold,JosefPerktoldMaintainersbashtagejosefpktdmatthew.brettrgommersSkipper.SeaboldClassifiersDevelopmentStatus4-BetaEnvironmentConsoleIntendedAudienceDevelopersScience/ResearchLicenseOSIApproved::BSDLicenseOperatingSystemOSIndependentProgrammingLanguagePython::2.6Python::2.7Python::3.2TopicScientific/EngineeringProjectdescriptionProjectdetailsReleasehistoryDownloadfilesProjectdescriptionWhatStatsmodelsisStatsmodelsisaPythonpackagethatprovidesacomplementtoscipyforstatisticalcomputationsincludingdescriptivestatisticsandestimationandinferenceforstatisticalmodels.MainFeatureslinearregressionmodels:Generalizedleastsquares(includingweightedleastsquaresandleastsquareswithautoregressiveerrors),ordinaryleastsquares.glm:Generalizedlinearmodelswithsupportforalloftheone-parameterexponentialfamilydistributions.discrete:regressionwithdiscretedependentvariables,includingLogit,Probit,MNLogit,Poisson,basedonmaximumlikelihoodestimatorsrlm:RobustlinearmodelswithsupportforseveralM-estimators.tsa:modelsfortimeseriesanalysis-univariatetimeseriesanalysis:AR,ARIMA-vectorautoregressivemodels,VARandstructuralVAR-descriptivestatisticsandprocessmodelsfortimeseriesanalysisnonparametric:(Univariate)kerneldensityestimatorsdatasets:Datasetstobedistributedandusedforexamplesandintesting.stats:awiderangeofstatisticaltests-diagnosticsandspecificationtests-goodness-of-fitandnormalitytests-functionsformultipletesting-var