IBM SPSS Statistics 26.0 IF006 Win/Linux/Macosx   專業的數據統計分析軟件
專業的數據統計分析軟件，它的主要功能就是數據錄入、資料編輯、數據管理、統計分析、報表製作、圖形繪製等等，並且軟件本身以及功能方面也還在不斷完善、更新。同時使用操作也是非常簡單，非常容易上手

IBMSPSSStatistics26.0IF006| 
Languages:English,Português,Français,Deutsch,Italiano,日本語,한국어,Polski,
Русский,Español,Simplified中文,Thaditional中文.
InAprilofthisyear,IBMreleasedthelatestversionofSPSSStatistics.Version26introducesanumberofadditionalanalysisproceduresaswellasnewcommandenhancements.

What’snewinIBMSPSSStatisticsv26

ThispagepresentsabriefoverviewofkeynewfeaturesinSPSSv26.Thesefeaturesbringmuchdesirednewstatisticaltests,enhancementstoexistingstatisticsandscriptingprocedures,andnewproductionfacilitycapabilitiestotheClassicuserinterface,whichalloriginatedfromcustomerfeedback.Hereiswhatyoucandowiththenewfeaturesinversion26:

Newanalyticalprocedures

QuantileRegression 
Instandard‘leastsquares’regressionthemodelpredictionsarebasedonasingleregressionline.Thislinecanbeusedtoestimatethemeanvalueofthedependentvariableasrepresentedbythepointsclusteringaboutlineatagivenvalueoftheindependent(predictor)variable(seefigure1)

Youmaynotefromthechartthatthereseemstobeaslight‘funnelling’ofthepointsnearthehighervaluesinthescatterplot.Technically,thisisreferredtoas‘heteroscedasticity’,butmoreprosaically,itjustindicatesthatthemodelislikelytobeworseatestimatinghighervaluesthanloweronessincethepointsvarymoreabouttheline.Quantileregressionoffersustheopportunitytofitthemodelusingamedianvalueratherthanamean.Weshouldbearinmindthatamedianisalsocalledthe50thpercentileandinthiscontextpercentileandquantilerefertothesamething.Althoughthere’snoreasontobelievethataregressionbasedonlinefittedaboutthemedianwouldbemoreaccuratethanonebasedonamean,quantileregressionisflexibleenoughtoallowustofitamodelbasedonotherpercentilevalues.Inotherwords,wecanfitseparateregressionlinesfordifferentpercentiles.Forexample,wecanrequestestimatesforthelowest10percent(quantile=0.1)orthetop90percent(quantile=0.9)ofthedependentvariable.

Theeffectofthisisthatwecanproduceseparatepredictionsforthedifferentpartsofthedependentvariable’sdistribution.UsingstandardLinearRegressiononthesamedatasetwegetasingleformulaforestimatingarespondent’scurrentsalary.Thisformulaconsistsofasinglecoefficientof1.9,meaningthatforeveryextradollarofbeginningsalary,therespondentearns$1.9dollarsintheircurrentsalary.Theformulaalsocontainsaconstantvalue(orintercept)of$1,928.However,there’snoreasontoassumethatthesameformulaappliestothedatainthetop10%ofthecurrentsalarydistribution,orthesay,thebottom25%.Assuch,Quantileregressionproducesseparatecoefficientsandinterceptvaluesforeachrequestedquantile.ThenewQuantileregressionprocedureevenplotsthesevaluesasshowninfigure3.

Thechartsevenshowtheparametervaluesforastandard(OLS)linearregressionmodelforcomparison(asindicatedbytheredline).Figure3illustratesthatnotonlydowegetdifferentinterceptvaluesfordatainthe20thpercentile(quantile0.2)vsthe80thpercentile(quantile0.8),butwealsogetdifferentparameterestimatesforthecoefficientvalues.

ROCAnalysis

ThenewROCproceduremakesiteasiertoassesstheaccuracyandperformanceofpredictiveclassificationmodels.ROC(ReceiverOperatorCharacteristic)analysisisspecificallyconcernedwiththeclassificationaccuracyofmodels,especiallyasregardstherelationshipsbetweentheaccurateclassifications(knownastheTruePositivesandTrueNegatives)andtheinaccuratepredictions(theFalsePositivesandFalseNegatives).TheseareoftenrepresentedbyaROCcurvethatplotsthetruepositiverate(TPR)againstthefalsepositiverate(FPR)atvariousthresholdsettings.ThenewROCAnalysisprocedurealsoincludesprecision-recall(PR)curvesandprovidesoptionsforcomparingtwoROCcurvesthataregeneratedfromeitherindependentgroupsorpairedsubjects.

BayesianStatistics
SPSSStatisticsv26alsoincludesenhancementstoitssuiteofbayesianstatisticalprocedures.
One-wayRepeatedMeasuresANOVA
TherepeatedmeasuresenhancementallowstheanalysttoadoptaBayesianapproachtocomparinganychangesinagivenfactorforthesamesubjectatdifferenttimepointsorconditions.Itisassumedthateachsubjecthasasingleobservationforeachtimepoint.
OneSampleBinomialenhancements
HeretheusermayapplyaBayesianbinomialtesttoattempttodeterminethelikelihoodthattheobservedratiobetweentwogroupsarethesameasanassumedproportioninthepopulation.
OneSamplePoissonenhancements
Likethepreviousprocedure,exceptheretheusermaycomparetheirdatatohowwellitfitsaPoissondistribution.Thesedistributionsareausefulmodellingforrareeventssuchasaccidentsorinsuranceclaims.AconjugatepriorwithintheGammadistributionfamilyisusedwhendrawingBayesianstatisticalinferenceonPoissondistribution.
ReliabilityAnalysis
SomeadditionalenhancementshavebeenmadetoSPSSStatistics’reliabilityprocedures.ReliabilityanalysishasnowbeenupdatedtoprovideoptionsforFleiss’MultipleRaterKappastatistics.Thistechniqueisoftenemployedwhenassessingthereliabilityofagreementbetweenafixednumberofraterswhenassigningcategoricalratingstoanumberofitemsorclassifyingitems.Thiscontrastswithotherkappavalues(suchasCohen’skappa)whichonlyapplytoassessmentsofagreementbetweenamaximumoftworaters.

TheIBMSPSSsoftwareplatformoffersadvancedstatisticalanalysis,avastlibraryofmachine-learningalgorithms,textanalysis,open-sourceextensibility,integrationwithbigdataandseamlessdeploymentintoapplications.Itseaseofuse,flexibilityandscalabilitymakeSPSSaccessibletouserswithallskilllevelsandoutfitsprojectsofallsizesandcomplexitytohelpyouandyourorganizationfindnewopportunities,improveefficiencyandminimizerisk.

Version26introducesanumberofadditionalanalysisproceduresaswellasnewcommandenhancements.

IBM(InternationalBusinessMachines)ranksamongtheworld'slargestinformationtechnologycompanies,providingawidespectrumofhardware,softwareandservicesofferings.

Product:IBMSPSSStatistics
Version:26.0IF006*
SupportedArchitectures:32bit/64bit
WebsiteHomePage:http://www.ibm.com
Language:multilanguage
SystemRequirements:**
SupportedOperatingSystems:**