應(yīng)用時(shí)間序列分析第4章答案

1、河南大學(xué):姓名：汪寶班級(jí)：七班學(xué)號(hào):1122314451班級(jí)序號(hào):685:我國(guó)1949年2008年年末人口總數(shù)（單位：萬(wàn)人）序列如表4-8所示（行數(shù)據(jù)）.選擇適當(dāng)?shù)哪Ｐ蛿M合該序列的長(zhǎng)期數(shù)據(jù)，并作5期預(yù)測(cè)。解：具體解題過(guò)程如下：（本題代碼我是做一問(wèn)寫一問(wèn)的）1：觀察時(shí)序圖：datawangbao4_5;inputx;time=1949+n-1;cards;54167551965630057482587966026661465628286465365994672076620765859672956917270499725387454276368785348067182992852298717789

2、21190859924209371794974962599754298705100072101654103008104357105851107507109300111026112704114333115823117171118517119850121121122389123626124761125786126743127627128453129227129988130756131448132129132802;procgplotdata=wangbao45;plotx*time=1;symbol1c=blackv=stari=join;run;分析：通過(guò)時(shí)序圖，我可以發(fā)現(xiàn)我國(guó)1949年-200

3、8年年末人口總數(shù)(隨時(shí)間的變化呈現(xiàn)出線性變化.故此時(shí)我可以用線性模型擬合序列的發(fā)展.Xt=a+bt+Itt=1,2,3,60E(It)=0,var(It)=2其中，It為隨機(jī)波動(dòng)；X=a+b就是消除隨機(jī)波動(dòng)的影響之后該序列的長(zhǎng)期趨勢(shì)。2:進(jìn)行線性模型擬合：procautoregdata=wangbao4_5;modelx=time;outputout=outp=wangbao4_5_cup;run;procgplotdata=out;plotx*time=1wangbao4_5_cup*time=2/overlay;symbol2c=redv=nonei=joinw=2l=3;run;TheS

4、ASSystem15:Wednesday,December7,20131TheAUTDREEiProcedureDependentVariftblexOrdinaryLeastSquaresEstinatesSSE262286607DFt53MSE4522356MotMSE2127SBC35冊(cè)383AIC1031.71114RegressR-Square0.9931ToialR-Squ&neO.gqglCurbin-WatsonO,0?24StsirdardApprcxVariableDFEstlftErrortWluePr|t|Intercept1-2770828S136-88.84.000

5、1time1449凡8591邰 FModel?2.46 OSES1&04EB3BB7S.9*0001Errorm3956243165.QCorrected Total1272.464SESApproxParget erEst imate Std Error Approximate 9SX Donf Ide nee Limits604.e 13.0152B鞏11?.24J333104JE0307(L0UU?8Bl.nani680.?12P-?1.0312得到的擬合模型為：xt 604.8 112.2 1.0307tt t=1， 2,，128lieratIvePhaseRer金bcSum口ISqu

6、ares00.10080.10001.10009.521BEJB1112E-20.39581.06781L8825E821123.10.52481JG隔1，劇口5ES81119JI0.71481.06251L8S03E841114,0，麻145361L8S01EC511叱E1,41111.05661L8778E8610施81Q53511,就就Efi71090.32.69911.050510489E8；81079.04.13611,0476LS15E8910CE.75.6224I.Q4501L7E34E8101035.3110.25071.04031.7+E811975.821.07871.05

7、471.67920812796.6El.5305l.0S11L3238E8ISG34.LI101.8I.0S922284992714614.1107.01.031254300516GOS.9111.71.030785597516604.8112.21.030783562JMG08112.21.0307855G2J16即白112,21.0307版6MApprawilesiteCtrr&ktianM亂ti雙abcaLOOOOODOb0*8758873c0.8427588-0.87599790,8427689I.ODODOOO*0,8361736-n.93517801.00000q03、擬合效果為了

8、直觀看出擬合效果，我們可以將原序列值和擬合值聯(lián)合作圖:procgplotdata=out;plotx*t=1xhat*t=2/overlay;symbollc=blackv=stari=join;symbol2c=redv=nonei=join;分析：由上圖圖我們可以看出，原序列值和擬合值很接近，擬合效果較好。綜合以上的分析，我們可以選擇模型：Xt604.8112.21.0307tt來(lái)擬合該序列的長(zhǎng)期趨勢(shì)。擬合效果很不錯(cuò)。8:某城市1980年1月至1995年8月每月屠宰生豬的數(shù)量（單位：頭）如表411所示（行數(shù)據(jù)），選擇適當(dāng)?shù)哪Ｐ蛿M合該序列的發(fā)展，并預(yù)測(cè)1995年9月至1997年9月該城市的生

9、豬屠宰量。解：具體解題過(guò)程如下：（本題代碼我是做一問(wèn)寫一問(wèn)的）atawangbao48;inputx;time=n;cards;76378719473387396428105084957411106471003119413310305590595101457768898129191643962281027361002641034919702795240916801012591095647689285773952109377198202979061003069408910268077919935611170628122588357106175919221041141099599788010538

10、696479975801094901101919097498981107188941771150971136961145321201109360711092510331212018410306910335111133110616111159099447101987853338697010056189543892658271979498748467381977029784468697875878695717572264182773576329259380783327238155971697508547270133791258580581778868526906979556881746669872

11、2587344576131860827544373969781397864666269737768003470694818237564075540822297534577034785897976975982780747758884100979668905193503847477453191900-8163589797810227826577271850439541879568103283957709129710124411452510113993866951711001831039261026431083879707790901903368873283759992677329278943943

12、999293790130910551060621035601040751017839379110231382413835341090119649910243010300291815990671100671015999764610493088905899361067238430711489610674987892100506；procgplotdata=wangbao48;plotx*time=1;symbol1c=redi=joinv=star;run;procarimadata=wangbao4_8;identifyvar=x;run;1:時(shí)序圖與平穩(wěn)性判別moQO；i70 Bl 9D 1

13、加 ird12 口m Ida 150 ICO 1?O IO H9DIStOQl：1ICO0：IMOOG四：knooTbTb1rh0.1422527fc15?(J33bJ.4228b雷21附小巡指gU_149956979781?：3fi0,41353出來(lái)濯M席海噩棒0,166170S082Si880.41800水踹邛制灰踹耀0.16139110eeC510?70.8456251陋串出9棚WQ.1C75041175S111B90.39352和H幅0-171282I?536363ZIU.413tj4l富*技和1：雅米木幫出U.1I35713SS75D.29S2I*來(lái)濯I*.0.19300714572

14、040070.29S56季米邛和榴K.0,19654C1551274826%卬事聘詠.0.139044IB3?1189500.192420.19003217520857630.2TOO1端瞄呻世ts.0.1910661833414930.178030.193086分析：由上圖可知：樣本自相關(guān)圖中的自相關(guān)系數(shù)在延遲4階之后幾乎全部落入2個(gè)標(biāo)準(zhǔn)差范圍之內(nèi)，并且向零衰減的速度還是比較快的。所以我認(rèn)為該序列是平穩(wěn)的序列。由時(shí)序圖與自相關(guān)圖可知其是平穩(wěn)序列。故可以用第三章的AR莫型或MAI型或ARM模型進(jìn)行擬合Ihe占需Asternm加版由上Lecetiiber3TheARIMAProcedureAut

15、occrrelationCheckforWhiteNoise分析：由上面數(shù)據(jù)可知：由于p值顯著小于0.05 ,故可以否定原假設(shè)（H0）,接受備選假設(shè)（Hi）,即我可以認(rèn)為該ToLagEChiSquare311.38OF6Pr)ChiSq.o(ma._g_.0.5340.535huiuuurruii.iuii0.4860.46D0.685043012519班12J0D10.4230.4UD,4180.U60.384148518599JG18M0.2能。上97DJES0J320.2700.17824634.124：.00010.1790.182ft.1200J3fl0.1730.186精品資料序列

16、是平穩(wěn)的非純隨機(jī)性序列。這樣就說(shuō)明了我們可以根據(jù)歷史信息預(yù)測(cè)未來(lái)月的生豬屠宰量。PartialAutocorrelatIonsL岫CorrelationH99765432I0IZ345678910.59421*，出*犒卜湍點(diǎn)IHH(用Hi0.29098UJ曲 -0.063U,出帶*格爾iC.18283HtMdt0.03697申.0.08689出明0.01556D.1022C*-0.09136a0.1378?由耕0.20160-C.20333-0.09154-0,I1B45，榔0.004490.05737出.-0.09712:榔0.00789-0.08069-C.017D40.04719出：0.

17、0&090出.0.0854?庫(kù).1 nt UO456789 012345679901234,1 1 ,1 JI i i JI 1 JI 1 ,n/- 2 2 OL分析：觀察自相關(guān)圖和偏自相關(guān)圖，從這兩圖來(lái)看，偏自相關(guān)圖是拖尾，而自相關(guān)系數(shù)是拖尾的。因而我們可以用ARMAI型進(jìn)行擬合。但是為了穩(wěn)妥起見(jiàn)，我還需要利用計(jì)算機(jī)進(jìn)行相對(duì)最優(yōu)定階。2:相對(duì)最優(yōu)定階：二一identifyvar=xnlag=18minicp=(0:5)q=(0:5);run;17gd2Wfiarlrify.jsi-jsrnt-i-iarThvnijmImfpr*naticnnrItori后(舊舊Hl s71lniu&MINI

18、 -1 二二君6Bn BH1S旦lllnKrl B i日口7日 日曰z曰41- 22111 BABB8BGn BR/H 口 sBBHg 7003&1 7H211 . -n V - nA 8n a_H MlBR 7B3S Bn44bn-B7-4IS工 曰HNl : -Ms B _H B H11 nil14S746 H 了b后才fl 日rly丁 FFS B -Hill -H - - BVBB837619-41 7丁口* UN nlm 口 SM-S72S 日41 -1 -B -s - SHInlsnlzsIt|LaeMU7S42E.570G4.410.fi?G90D11MAIJ0.844110.23

19、8953.680.90D51仙1,2-0.480880J3638-2.C20.14512NAI,3-0.424?0J3831-M3(1.13463MAIH0.G8313O.095G1C.54fl65LOO!-D.61?明，4-0.036-0.0970.366-0.61Z1.000Ml,5-0.021-0.791口5g-0.36-0.26?軀1J-o.oo?U.脂。-0.S750.7540,19?陽(yáng)，2-0,016-0.S950.934-0.8190.202NR14蚓.40,007-0.0240.7S6-0.87CQ.SS7-0.9(11OJ9-O.fi75-k4co0.3Q3Correlati

20、onsofParameterEatimatesParameterL1AU5幅1,1ARL2AAUARL4MU-0.0210.002-0.018D.0?-0.024蜘-0.7910.350-0.895QJ86-OJ01MAU9.580-9760.934-0.S7EMR,3MAI,4-9J640e?54-U.8130.897-04875-0.1920.282-0.1000.393MAL6ARIJi.ooa-0.717Q.641-0.513-0,71?1.D0D-0,S510.84Q-Q,044【he$福%；43Wednesday,DeGLeri,ZUS6TheARIMAProcedureCorre

21、lationsofP社r。幀terEstin)3itea6 2 8 40-61 1 3 3088131 o- TJ1 j 1 1 2 3 4 51-0.003。,琢-0.0233QJW5-0.1Q3-0.031Q ,州2-Q.Q269口.叩44-0,0730.30.020-0.084150,0084-0.0270刃28-0.)63(M0521O.OOG7-0.029口，014-0.043-0.0960.00330.0840,062-0.136Mede I for var iabl e xEst iimated Mean 78426.477 8807 M 4 2- 5 0- 3 4 nuoooo

22、o o- Au nu o4 84 6 06 D 5 6 s2- 1 1 1 oRarameterMAU5ARIJARH2ARlf4峭1J口.641-0.9511.000-0.334O.fi57同，9-0,5190.940-0.954L010-0.942期1,4U.56S-0.84-40.057-0.3421.000AutoccrrelaiIanChecl(ofResiduaIsToChi-PrLacSiusreDFChiSqflutocorrelationsftutoregressiveFactorsFactor1：1-L21457+0.70228Q.04985B*(3)-0.41243肺1該

23、輸出形式等價(jià)于Xt=(1-1.21457B+0.70228B2-0.04985B3-0.41243B4)?t故該模型為：Xt=+(0(B)/MB)?t=(3)序列預(yù)測(cè)(1995年9月至1997年9月)forecastlead=24id=timeout=wangbaoyc;|run;TheSASSyslen16:43Wednesdiiy.Deceoiber7,201317TheARIMAProcedureForecastsforvariablexObsFmecu乳StdError551ConfidanceLimits901245 6 7 Gogol 2 34-07008012 de _9 9 -

24、3 3 fir o Av o n- _u _u -u i JI 1 11II111111 1- 2 2 222222222229C315.9034 96460.1492 9C995.9518 97686.9964 9B798.995S8E466.8020 948E9,7G2E 95182.1334 9559LUE5U 3531?.2504 94433.7554 8373?.4261 的附加3 93868.2094 83812.923B 93290.OC9C 32B67.5393 9298 9. GO 4E82335.8456 92990.4099 32083.B367 91615.72S4 91253.8459 91133.909958.1433 10618.509 1OHO.049 11913.376 12O8R.0O5 12386.880 12765.493 13342.391 1381g.2兆 15.建1 1423.480 14573.44614931.425 1594JOO 15567.402 157G9.222 15954.381 1S199.221 16468.E27 18693.060 16878.688 17033.535 17206.90417406.11G76790.3013 75646.2941 75709.0769 ?420f