practical regression and anova using r

《practical regression and anova using r》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。

1、PracticalRegressionandAnovausingRJulianJ.FarawayJuly20021Copyrightc1999,2000,2002JulianJ.FarawayPermissiontoreproduceindividualcopiesofthisbookforpersonaluseisgranted.Multiplecopiesmaybecreatedfornonprofitacademicpurposes—anominalchargetocovertheexpenseofreproductionmaybemade.Reproducti

2、onforprofitisprohibitedwithoutpermission.PrefaceTherearemanybooksonregressionandanalysisofvariance.Thesebooksexpectdifferentlevelsofpre-parednessandplacedifferentemphasesonthematerial.Thisbookisnotintroductory.Itpresumessomeknowledgeofbasicstatisticaltheoryandpractice.Studentsareexpected

3、toknowtheessentialsofstatisticalinferencelikeestimation,hypothesistestingandconfidenceintervals.Abasicknowledgeofdataanalysisispresumed.Somelinearalgebraandcalculusisalsorequired.Theemphasisofthistextisonthepracticeofregressionandanalysisofvariance.Theobjectiveistolearnwhatmethodsareavai

4、lableandmoreimportantly,whentheyshouldbeapplied.Manyexamplesarepresentedtoclarifytheuseofthetechniquesandtodemonstratewhatconclusionscanbemade.Thereisrelativelylessemphasisonmathematicaltheory,partlybecausesomepriorknowledgeisassumedandpartlybecausetheissuesarebettertackledelsewhere.The

5、oryisimportantbecauseitguidestheapproachwetake.Itakeawiderviewofstatisticaltheory.Itisnotjusttheformaltheorems.QualitativestatisticalconceptsarejustasimportantinStatisticsbecausetheseenableustoactuallydoitratherthanjusttalkaboutit.Thesequalitativeprinciplesarehardertolearnbecausetheyare

6、difficulttostatepreciselybuttheyguidethesuccessfulexperiencedStatistician.Dataanalysiscannotbelearntwithoutactuallydoingit.Thismeansusingastatisticalcomputingpack-age.Thereisawidechoiceofsuchpackages.Theyaredesignedfordifferentaudiencesandhavedifferentstrengthsandweaknesses.Ihavechosento

7、useR(ref.IhakaandGentleman(1996)).WhydoIuseR?Theareseveralreasons.1.Versatility.Risaalsoaprogramminglanguage,soIamnotlimitedbytheproceduresthatarepreprogrammedbyapackage.ItisrelativelyeasytoprogramnewmethodsinR.2.Interactivity.Dataanalysisisinherentlyinteractive.Someolderstatis