統(tǒng)計專業(yè)英語教案 常用統(tǒng)計學術(shù)語復習

一、統(tǒng)計學術(shù)語

population總體

sample樣本

census普查

sampling抽樣

quantitative量的

qualitative質(zhì)的

discrete離散的

continuous連續(xù)的

populationparameters總體參數(shù)

samplestatistics樣本統(tǒng)計量

descriptivestatistics敘述統(tǒng)計學

抽樣調(diào)查samplingsurvey

簡單隨機抽樣simplerandomsampling

系統(tǒng)抽樣systematicsampling

分層抽樣stratifiedsampling

整群抽樣clustersampling

多級抽樣multistagesampling

實驗設(shè)計DesignofExperiment)

參數(shù)Parameter

Statistics統(tǒng)計學

Statisticaltable統(tǒng)計表

Statisticalchart統(tǒng)計圖

Piechart圓餅圖

Stem-and-leafdisplay莖葉圖

Histogram直方圖

BarChart長條圖

Polygon多邊形

Expectation期望值

Mode眾數(shù)

Mean平均數(shù)

Variance變異數(shù)

Standarddeviation標準差

Standarderror標準誤

Inferentialstatistics推論統(tǒng)計學

Pointestimation點估計

Intervalestimation區(qū)間估計

Confidenceinterval置信區(qū)間

Confidencecoefficient置信系數(shù)

Regressionanalysis回歸分析

Analysisofvariance變異數(shù)分析

Correlationcoefficient相關(guān)系數(shù)

Reliability信度

Validity效度

Discreteuniformdensities離散的均勻密度

Binomialdensities二項密度

Hypergeometricdensities超幾何密度

Poissondensities卜松密度

Geometricdensities幾何密度

Negativebinomialdensities負二項密度

Continuousuniformdensities連續(xù)均勻密度

Normaldensities正態(tài)密度（分布）

Exponentialdensities指數(shù)密度

Gammadensities伽瑪密度

Betadensities貝他密度

Multivariateanalysis多變量分析

Principalcomponents主因子分析

Discriminationanalysis判另U分析

Clusteranalysis群集分析

Factoranalysis因素分析

Survivalanalysis存活分析

Timeseriesanalysis時間序列分析

Linearmodels線性模式

Probabilitytheory概率率論

Statisticalinference統(tǒng)計推論

Stochasticprocesses隨機過程

Decisiontheory決策理論

Discreteanalysis離散分析

Mathematicalstatistics數(shù)理統(tǒng)計

相關(guān)系數(shù):Correlationcoefficient

算術(shù)平均數(shù)(ArithmeticMean)

元素(Element)郵寄問卷法(MailInterview)

封閉式問題(CloseQuestion)

電話訪問法(TelephoneInterview)

市場調(diào)查(MarketingResearch)

決策樹(DecisionTrees)

容忍誤差(Toleratederro)

數(shù)據(jù)挖掘(DataMining)

初級資料(PrimaryData)

趨勢分析(TrendAnalysis)

神經(jīng)網(wǎng)絡(NeuralNetwork)

人員訪問法(Interview)判別分析法(DiscriminantAnalysis)

集群分析法(clusteranalysis)規(guī)則歸納法(RulesInduction)

內(nèi)容效度(ContentValidity)判斷抽樣(JudgmentSampling)

二、閱讀資料

Frequency

Whencollectinginformation,forinstancethecolorofcarsinacar

park,therewillberepeatedexamplesofparticularcolors.Theremaybe

fouryellowcars,13redcars,eightbluecarsand20carsofothercolors.

Theinformationisqualitative.Thenumberofcarsofeachcoloristhe

frequencyofobservationofthatitemofinformation.

Inevitableingatheringanyinformationtherewillbeacollectionof

frequenciesassociatedwithitemsofdata.Thefrequenciesarenotonly

thedata,theytellussomethingaboutthedistributionofthedata.

Presentationofdata(1)

Qualitativedatamaybepresentedinafrequencytablesuchasthe

onebelow.

Amongthemanyconsiderations,itwouldbeimportanttohavesome

ideasofwhatatypicaltimewaslikelytobe,andthelikelyrangeoftimes.

Thisinformationisnecessarysothatanappropriatetimingdevicecanbe

selected.Therewouldbelittlepointintryingtouseawatch'sminute

handoranelectronictimercapableofrecordingto1/1000ofasecond.It

isalsounlikelytobenecessarytorecordtimesaslongasanhour,butthe

timerneedstobecapableofrecordingmorethanafewseconds.

Therearetwomeasuresrepresentativeofdatawhichmaybeuseful

incaselikethis-onerepresentingthe'typicalvalue5andanotherwhich

indicatesthe'sortofrangeofvalue5likelytobefound.Instatisticalterms,

thesearemeasuresoflocationoraverageanddispersionorspread.

Lifewouldbesimpleiftherewerejustoneofeach,orevenanideal

measureofeachquality.Unfortunately,thisisnotthecase.Thereare

manytypesofaverageandseveralkindsofspread.Inthischapterwe

shallconcentrateonaverages.

Mode

Themodeisthemostcommonlyoccurringvalueoritemofdate,or,

inotherwords,theonethatappearsmostfrequently.Inthecontextofthe

dataunderconsideration,themostcommonlyoccurringvalueis9

seconds.Isitreasonable,though,toconsider9secondsasbeingtypical

ofthetimetakentopassthroughthiscontinentalmotorwaytoll?

Almostcertainlynot!Itmanybemoreappropriatetoconsiderthemodal

class.Referringbacktothediagram,the1-branchhasthegreatest

frequency.Hence,itwouldbereasonabletosaythatthemodeisthetime

between20and30seconds.Thisisthelongestbranchinthe

stem-and-leafdiagram.

Themodalclassmaybetheclasswiththehighestfrequencywhen

thedataarepresentedinafrequencytable,butitmaynot!

Theprominenceofthe20-30secondsclassisapparent.Themethod

assumesthatthemodedividesthemodalclassinthesameratioasthe

increaseinfrequencydensitytothedecreaseinfrequencydensity.Inthe

frequencytable,thisratiois(9-7):(9-6),whichisequivalentto2:3.Hence

themodedividesthemodalclassintheratio2:3,andanestimateofthe

modeis24.Weneedtoaskourselves,howvalidisthisprocess?

WhereWisthewidthofthemodalclass,andxisitslowerbound.

Intheexample,1=2,D=3,W=10andx=20.

Hereestimateofthemode=20+(2/5)*10=24.

Median

Step-by-step

Thecentreormiddleitemofthedataisknownasthemedian.One

approachtoidentifyingthemedianisto:

■placethedatainorder

■locatethemiddleitem

■hence,identifythemedian

Supposeweneedtoidentifythemedianofthefollowingcollection

ofdata.

8,15,7,10,4,3,8,6,5,7,8

Placingthedatainorderyields:

3,4,5,6,7,7,8,8,8,10,15

Themiddleitemistheonewhichisequidistantfromtheextreme

values.

Sincethereareelevenitemsofdata,themiddleisthesixthfrom

eitherend.

Eventotal

Forthecollectionabove,thetotalnumberofitemsisodd,whichled

tothemedianbeingoneoftheactualrecordeditemsofdata.Inthe

followingcase,thetotalisanevennumber,whichmeansthecentervalue

ofthedataismidwaybetweentwooftherecordeditems.

4,5,0,3,9,4,8,9,9,1

Orderingthedatagives:

0,1,3,4,4,5,8,9,9,9

andlocatingthemiddlevalueyields:

Groupedfrequencytable

Ifthedataarepresentedinafrequencytable,thenitisonlypossible

toobtainanestimateofthemedian.Thisisdoneeithergraphicallyor

arithmetically.

Thecumulativefrequenciesforthefrequencytableisgivenbelow.

TimeFrequencyCumulativefrequency

[0,10)44

[10,20)711

[20,30)920

[30,40)626

[40,50)531

[50,60)334

[60,120)943

Groupedfrequencytable

Whatinformationcanbegainedfromthecumulative

frequency?

Considerthecumulativefrequencyof20.thistellsusthatthereare

20itemsofdatawithvalueslessthan30.Similarly,thecumulative

frequencyof34indicatesthatthereare34itemswhicharelessthan60.

Thereisanaturallinkbetweenanygivencumulativefrequencyandthe

upperboundofthecorrespondingclass.Hence,whenitcomesto

constructingacumulativefrequencygraph,thepointstobeplottedcome

fromthefollowingseriesdata.

Range

Perhapsthemostsimplemeasureofspreadisthedifferencebetween

thelargestandsmallestitemsofdatai.e.thedifferencebetweenthe

extremes.Thisistherange.Inthecaseofthemotorwaytolltimesthe

longesttimerecordedwas118secondsandtheshortesttimewas9

seconds,hencetherangeofthesedataisgivenby:

range=118-9-109seconds

Thismeasureofspreaddosenottakeintoaccountanythingabout

thedistributionofthedataotherthantheextremes.Neitherisitvery

reliableortypical.Why?

Quartilespread

Amoretrustworthymeasureistherangeofthemiddlehalfofthe

data.Toidentifythisrangeweneedtofindtheitemsofdatawhichare

positionedhalfwaybetweentheextremesandthemedian.Takethecase

ofthedatafollowing:

Ingeneral,theitemsofdatalyingmidwaybetweenthemedianand

theextremesareknowasthequartiles.Itismoreusualtorefertothemas

thefirstorlowerquartileandthethirdorupperquartile.Thedifference

betweenthemiscalledtheinterquartilerange(IQR)orquartilespread

(QS)

Standarddeviation

Theinterquartilerangemeasuresthespreadofthemiddlehalfofthe

dataandiscloselylinkedtothemedian.Wecandefineameasureof

dispersion,takingintoaccountallthedata,whichislinkedinsteadtothe

mean.

Deviationfromthemean

Supposetheitemsofdata:18,20,21,22,24.

Themeanofthecollectiondatais21,hencethedeviationfromthe

meanare:-3,-1,0,1,3,andtheaverage(ormean)ofthisdeviationsis

theirsumdividedbythenumberofitems.Thisc