PubH 7405: REGRESSION ANALYSIS DIAGNOSTICS IN MULTIPLE REGRESSION

Size: px
Start display at page:

Download "PubH 7405: REGRESSION ANALYSIS DIAGNOSTICS IN MULTIPLE REGRESSION"

Transcription

1 PubH 7405: REGRESSION ANALYSIS DIAGNOSTICS IN MULTIPLE REGRESSION

2 The daa are n he form : {( y ; x, x,, x k )},, n Mulple Regresson Model : Y β 0 β x β x β k x k ε ε N (0, σ ) The error erms are dencally and ndependenly dsrbued as normal wh a consan varance.

3 A smple sraegy for he buldng of a regresson model consss of some, mos, or all of he followng fve seps or phases: () Daa collecon and preparaon, () Prelmnary model nvesgaon, (3) Reducon of he predcor varables, (4) Model refnemen and selecon, and (5) Model valdaon Le say we fnshed sep #3; we enavely have a good model: s me for some fne unng!

4 Dagnoscs o denfy possble volaons of model s assumpons are ofen focused on hese major ssues : () Non-lneary, () Oulyng & nfluenal cases, (3) Mul-collneary, (4) Non-consan varance & (5) Non-ndependen errors.

5 DETECTION OF NON-LINEARITY

6 A lmaon of he usual resdual plos (say, resduals agans values of a predcor varable): hey may no show he naure of he addonal conrbuon of a predcor varable o hose by oher varables already n he model. For example, we consder a mulple regresson model wh ndependen varables X and X ; s he relaonshp beween Y and X lnear? Added-varable plos (also called paral regresson plos or adjused varable plos ) are refned resdual plos ha provde graphc nformaon abou he margnal mporance of a predcor varable gven he oher varables already n he model.

7 ADDED-VARIABLE PLOTS In an added-varable plo, boh he response varable Y and he predcor varable under nvesgaon (say, X ) are boh regressed agans he oher predcor varables already n he regresson model and he resduals are obaned for each. These wo ses of resduals reflec he par of each (Y and X ) ha s no lnearly assocaed wh he oher predcor varables. The plo of one se of resduals agans he oher se would show he margnal conrbuon of he canddae predcor n reducng he resdual varably as well as he nformaon abou he naure of s margnal conrbuon.

8 A SIMPLE & SPECIFIC EXAMPLE Consder he case n whch we already have a regresson model of Y on predcor varable X and s now consderng f we should add X no he model (f we do, we would have a mulple regresson model of Y on (X,X )). In order o decde, we nvesgae smple lnear regresson models: (a) The regresson of Y on X and (b) The regresson of X on X and oban ses of resduals as follows:

9 Regresson# : Y on X ^ Y e b 0 ( Y X (frs or old b ) X Y ^ Y model) These resduals represen he par of Y no explaned by X

10 Regresson# : X ^ X e ( X on X b * 0 X (second ) b * X X or new model) ^ X These resduals represen he par of X no conaned n X

11 We now do a regresson of e (Y X ) as new dependen varable on e (X X ) as ndependen varable: Tha s o see f he par of X no conaned n X can furher explaned he par of Y no explaned by X ; (f can, X should be added o he model for Y whch already has X n ). There are hree possbles:

12 The horzonal band shows ha X conans no addonal nformaon useful for he predcon of Y beyond ha conaned n and provded for by X

13 Ths lnear band wh a non-zero slope ndcaes ha he par of X no conaned n X s lnearly relaed o he par of Y no explaned by X alone. Tha s, X should be added o form a -varable model. Noe ha f we do a regresson hrough he orgn, he slope b should be he same as he coeffcen of X f s added o he regresson model already conanng X.

14 Suppose : [ Y ^ () Y ( X ) X ^ () X ( X ) X and : (3) e( Y Then : X ) ( X AN EXAMPLE 6.9e( X )] 6.9[ X we have he same resul as n mulple regresson fng : Y [50.70 (6.9)(40.78)] 6.9X Y X 4.7X X ) ( X [5.54 (6.9)(.7)] X )],

15 The curvlnear band ndcaes ha, lke he case (b) prevously, X should be added o he model already conanng X. However, furher suggess ha he ncluson of X s jusfed bu some power erms or some ype of daa ransformaon are needed.

16 The fac ha an added-varable plo may sugges he naure of he funconal form n whch a predcor varable should be added o he regresson model s more mporan han ha he varable possble ncluson (whch can be resolved whou graphcal help). The added-varable plos play he role ha we use scaer dagrams for n smple lnear regresson; hey would ell f daa ransformaon or f ceran polynomal model s desrable.

17 ADDED-VARIABLE PLOT PROC REG daa Orgse; Smple Lnear Regresson here Model Y X X/noprn; Oupu Ou SaveMe R YResd XResd; Run; PROC PLOT daa SaveMe; Plo YResd*XResd; Run;

18 Anoher Example: Suppose we have a dependen varable Y and 5 predcor varables X-X5; and assume ha we focus on X5, our prmary predcor varable: () we regress Y on X-X4 and oban resduals (say, YR), and () we regress X5 on X-X4 and oban resduals (say, X5R), hen (3) we regress YR on X5R (SLR, no nercep) lnear assumpon abou X5 can be suded from hs las SLR & s (added-value) plo.

19 MARGINAL SL REGRESSION PROC REG daa Orgse; Model Y X5 X X X3 X4/noprn; Oupu Ou SaveMe R YR X5R; Run; PROC PLOT daa SaveMe; Plo YR*X5R; Run; PROC REG daa SaveMe; Model YR X5R/ non; Run; Mulple Lnear Regresson here New: Margnal Smple Lnear Regresson

20 REMEDIAL MEASURES If a lnear model s found no approprae for he added value of ceran predcor, here are wo choces: () Add n a power erms (quadrac), or () Use some ransformaon on he daa o creae a f for he ransformed daa

21 Each has advanages & dsadvanages: frs approach (quadrac model) may yeld beer nsghs bu may lead o more echncal dffcules; ransformaons (log, recprocal, ec ) are more smple bu may obscure he fundamenal real relaonshp beween Y and ha predcor. One s hard o do and one s hard o explan

22 OUTLYING & INFLUENTIAL CASES

23 SEMI-STUDENTIZED RESIDUALS _ e e e * MSE e MSE If MSE were an esmae of he sandard devaon of he resdual e, we would call e* a sudenzed (or sandardzed) resdual. However, he sandard devaon of he resdual s complcaed and vares for dfferen resduals, and MSE s only an approxmaon. Therefore, e* s call a sem-sudenzed resdual. Bu exac sandard devaons can be found!

24 THE HAT MATRIX ^ Y Xb X[(X' X) X' Y] [X(X' X) X']Y HY H X(X' X) X' s called he "Ha Marx" The Ha marx can be obaned from daa

25 RESIDUALS Model : Y Fed Value : ^ Y Resduals : e Xβ ε Xb ^ Y Y Y Xb Y HY (I H)Y Lke he Ha Marx H, (I-H) s symmerc & dempoen

26 VARIANCE OF RESIDUALS H) (I H) (I H) H)(I (I H) H)(I (I H)' I)(I H)( (I H)' (Y)(I H)σ (I (e) σ H)Y (I e MSE ^ ' σ σ σ σ

27 For he h observao n : σ s σ ( e s( e h h j (e ( e, e, e ) ) ) ( h ) ( h j j h h j j )σ ) MSE σ ; j MSE; j s he h elemen on he man s on h row and jh column of dagonal & he ha marx

28 STUDENTIZED RESIDUALS r e s( e ) ( e h ) MSE Sudenzed resduals s a refne verson of he semsudenzed resduals; n r we use he exac sandard devaon of he resdual e no an approxmaon. Sem-sudenzed resduals and Sudenzed Resduals are ools for deecng Oulers.

29 DELETED RESIDUALS The second refnemen o make resduals more effecve for deecng oulyng or exreme observaons s o measure he h resdual when he fed regresson s based on all cases excep he h case (smlar o he concep of jackknfng). The reason s o avod he nfluence of he h case especally f s an oulyng observaon - on he fed value. If he h case s an oulyng observaon, s excluson wll make he deleed resdual larger and, herefore, more lkely o confrm s oulyng saus: d Y ^ Y ( )

30 STUDENTIZED DELETED RESIDUALS Combnng he wo refnemens, he sudenzed resdual and he deleed resdual, we ge he Sudenzed Deleed Resdual. A sudenzed resdual s also called an nernal sudenzed resdual and a sudenzed deleed resdual s an exernal sudenzed resdual; MSE () s he mean square error when he h case s omed n fng he regresson model. d s( d ) ( h e ) MSE ( )

31 e ( h ) MSE( ) The sudenzed deleed resduals can be calculaed whou havng o f new regresson funcons each me a dfferen case s omed; ha s, we need o f he model only once wh all n cases o oban MSE no he same model n mes, here p s he number of parameers, p k: ( n p) MSE ( n p ) MSE ( ) e h

32 ( n leadng p) MSE o : n p e SSE( h) e whch s dsrbue d as "" wh degreesof ( n p ) MSE freedom under H ( ) 0 e h of (n p ) no oulers. Ths can be used as a sascal es for oulers wh allowance for mulple decsons

33 The dagonal elemens h of he ha marx, called leverages, have he same ype of nfluence on he sudenzed deleed resduals and, herefore, serve as good ndcaors n denfyng oulyng X observaons; he larger hs value he more lkely ha he case s an oulers. A leverage value s usually consdered o be large f s more han wce as large as he mean leverage value (whch s p/n).

34 "no good" s so large ; ) ( ), dempoen s ( ) symmerc, s ( ) ( ^ h h Y σ σ σ σ σ H HH H H H' H H HH H' I H (Y)H' Hσ (Y) σ HY Y ^ ^

35 IDENTIFICATION OF INFLUENTIAL CASES Wha we have done s o denfy cases ha are oulyng wh respec o her Y values (say, usng her sudenzed deleed resduals) and/or X values (usng her leverages). The nex sep s o asceran wheher or no hese oulyng cases are nfluenal; case s nfluenal f s excluson causes major changes n he fed regresson funcon. (If a case s deermned o be oulyng and nfluenal, he nex sep would be nvesgang usng oher sources o see f should be aken ou).

36 INFLUENCE ON A FITTED VALUE Recall he dea of deleed resduals n whch we measure he h resdual from he fed value where regresson fng s based on all cases excep he h case so as o avod he nfluence of he h case self. We can use he very same dea o measure he nfluence of a case on s own fed value; ha s o measure he dfference ( DF ) beween he fed values when all daa are used o he fed value where regresson fng s based on all cases excep he h case; MSE() s he mean square error when he h case s omed ( DFFITS ) ^ ^ Y Y MSE ( ) ( ) h

37 INFLUENCE ON ALL FITTED VALUES Takng he same dea of deleng a case o nvesgae s nfluence bu, n conras o DDFITS whch consder he nfluence of a case on s own fed value, Cook s Dsance shows he effec of h case on all fed values. The denomnaor serves only as a sandardzed measure so as o reference Cook s Dsance o he F(p,n-p) percenles. D n j ^ (Y j ^ Y pmse j() )

38 INFLUENCE ON REGRESSION COEFFICIENTS Anoher measure of nfluence, DFBETAS, s defned smlar o DDFITS bu DFBETAS focuses on he effec of (deleng) h case on he values of all regresson coeffcens; n hs formula, c jj s he jh dagonal elemen of marx (X X) -. As a gudelne for denfyng nfluenal cases, a case s consdered nfluenal f he absolue value of a DFBETAS exceeds for medum daa ses and / n for larger ses: ( DFBETAS) j( ) b j b MSE j( ) ( ) c jj

39 THE ISSUE OF MULTICOLINEARITY

40 When predcor varables are hghly correlaed among hemselves we have mulcollneary : he esmaed regresson coeffcens end o have large sandard errors. Oulyng and nfluenal effecs are case-based whereas mulcollneary effecs are varable-based.

41 CONSEQUENCES OF MULTICOLLINEARITY Addng or deleng a predcor varable changes some regresson coeffcens subsanally. Esmaed sandard devaons of some regresson coeffcens are very large. The esmaed regresson coeffcens may no be sgnfcan even her presence mprove predcon.

42 INFORMAL DIAGNOSTICS Indcaons of he presence of mulcollneary are gven by he followng nformal dagnoscs: Large coeffcens of correlaon beween pars of predcor varables n marx r XX. Non-sgnfcan resuls n ndvdual ess some esmaed regresson coeffcens may even have wrong algebrac sgn. Large changes n esmaed regresson coeffcens when a predcor varable s added or deleed.

43 For he purpose of measurng and formally deecng he mpac of mulcollneary, s easer o work wh he sandardzed regresson model whch s obaned by ransformng all varables (Y and all X s) by means of he correlaon ransformaon. The esmaed coeffcens are now denoed by b* s; and here s no nercep.

44 Correlaon ransformaon: s x x x n x _ * Wh ransformed varables Y* s and all X* s, he resul s called he Sandardzed Regresson Model: * * * * * * * * 0 & ε β β β ε β β β β k k k k x x x Y x x x Y Resuls: YX XX * r r X' Y X) (X' b XX ' * r (X X) ) (b σ * * σ σ

45 σ (b * ) σ σ * ' (X X) * I s obvous ha he varances of esmaed regresson coeffcens depend on he correlaon marx r xx beween predcor varables. The jh dagonal elemen of he marx r xx - be denoed by (VIF) j and s called he varance nflaon facor ): r XX

46 VARIANCE INFLATION FACTORS Le R j be he coeffcen of mulple deermnaon when predcor varable X j s regressed on he oher predcor varables, hen we have more smple formulas for he varance nflaon facor and he varance of he esmaed regresson coeffcen b* j. These formulas ndcae ha: () If R j 0 (ha s X j s no lnearly relaed a all o he oher predcor varables), (VIF) j, () If R j 0, hen (VIF) j > ndcang an nflaed varance for b* j

47 ( VIF ) j ( R j ) σ ( b * j ) σ * R j

48 COMMON REMEDIAL MEASURES () The presence of mulcollneary ofen does no affec he usefulness of he model n makng predcons provded ha he values of he predcor varables for nferences are nended follow he same mulcollneary paern as seen n he daa. One smple remedal measure s o resrc nferences o arge subpopulaons havng he same paern of mulcollneary. () In polynomal regresson models, use cenered values for predcors

49 COMMON REMEDIAL MEASURES (3) One or more predcor varables may be dropped from model n order o lessen he degree of mulcollneary. Ths pracce presens an mporan problem: no nformaon s obaned abou he dropped predcor varables. (4) To add more cases ha would break he paern of mulcollneary; hs opon s no ofen avalable.

50 COMMON REMEDIAL MEASURES (5) In some economc sudes, s possble o esmae he regresson coeffcens for dfferen predcor varables from dfferen ses of daa; however, n oher felds, hs opon s no ofen avalable. (6) To form a compose ndex or ndces o represen he hghly correlaed predcor varables. Mehods such as prncpal componens can help, bu mplemenaon s no easy (7) Fnally, Rdge Regresson has proven o be useful o remedy mulcollneary problems; Rdge regresson s one of he advanced mehods whch modfes he mehod of leas squares o allow based bu more precse esmaors of he regresson coeffcens.

51 NON-CONSTANT VARIANCE

52 Graph of Resduals agans Predcor varable or agans he fed values s helpful o see f he varance of error erms are consan; f model fs, shows a band cenered around zero wh a consan wdh. Lack of f resul n a graph showng he resduals deparng from zeros n a sysemac fashon a megaphone shape. No new fancy mehod/graph are needed here!

53 Example: Resdual Plo for Non-consan Varance

54 I s more me-consumng o deec nonlneary, bu s more smple o fx : A log ransformaon of an X or addon of s quadrac erm would normally solve he problem on non-lneary. I s more smple o deec a non-consan varance; added-value plo s no needed. However, would be more dffcul o fx because, n order o change he varance of Y we o make a ransformaon on Y.

55 Transformaons of Y maybe helpful n reducng or elmnang unequal varances of he error erms. However, a ransformaons of Y also changes he regresson relaon/funcon. In many crcumsances an approprae lnear regresson relaonshp has been found bu he varances of he error erms are unequal; a ransformaon would make ha lnear relaonshp non-lnear whch s a more severe volaon. An alernave o daa ransformaons: whch are more dffcul o fnd - usng mehod weghed leas squares nsead of regular leas squares.

56 Wh he Weghed Leas Squares (WLS), esmaors for regresson coeffcens are obaned by mnmzng he quany Q w where w s a wegh (assocaed wh he error erm); seng he paral dervaves equal o zero o oban he normal equaons : Q w k ( β β X ) 0 w Y

57 The opmal choce for he wegh s he nverse of varance. For example, when sandard devaon s proporonal o X 5 (or varance s kx 5 ), we mnmze: ) ( X Y X Q β β

58 Le consder, n more deph, he generalzed mulple regresson model: And frs look a he case where error varances are known and hen relax hs unrealsc assumpon. ) (0, 0 k k N x x x Y σ ε ε β β β β

59 When error varances are known, esmaors for regresson coeffcens are obaned by mnmzng he quany Q w where w s a wegh (assocaed wh he error erm, opmal choce s nverse of he known varance); seng he paral dervaves equal o zero o oban he normal equaons. The weghed leas squares esmaors of he regresson coeffcens are unbased, conssen, and have mnmum varance among unbased lnear esmaors Q w w β β 0 X β X ) ( Y k k

60 Le he marx W be a dagonal marx conanng (X' WX)b b σ w (X' (b w w ) σ he weghs w WX) X' WY (X' X' WY WX) 's, we have Wh Ordnary Leas squares, W I.

61 If he error varances were known, he use of WLS would be sragh forward: easy and smple. The resulng esmaors exhb less varably han he ordnary leas squares esmaors. Unforunaely, s an realsc assumpon ha we know he error varances. We are forced o use esmaes of he varances and perform Weghed Leas squares esmaon usng hese esmaed varances.

62 The process o esmae error varances s raher edous and can be summarzed as follows: () F he model by un-weghed (or ordnary) leas squares and obaned resduals, () Regress he squared resdual agans he predcor varables o oban a varance funcon (varance as a funcon of all predcors) (3) Use he fed values from he esmaed varance o oban he wegh (nverse of esmaed varance) (4) Perform WLS usng esmaed weghs.

63 A SIMPLE SAS PROGRAM If error varances are known, he WEIGHT saemen (no opon ) allow users o specfy he varable o use as he wegh n he weghed leas squares procedure: PROC REG; WEIGHT W; RUN; MODEL Y X X X3 X4; If error varances are unknown, would ake a few sep o esmae hem before you can use he WEIGHT saemen.

64 A SAMPLE OF SAS FOR WLS proc REG daa SURV; model SurTme LTes ETes PIndex Clong; oupu outemp RSurTLSR; run; daa TEMP; se Temp; sqrsurtlsr*surtlsr; run; proc reg daatemp; model sqrltes ETes PIndex Clong; oupu ou emp3 PEsqr; run; daa Temp4; se emp3; w/esqr; run; proc reg daaemp4; wegh w; model SurTme LTes ETes PIndex Clong; run; To form Varance Funcon

65 The condon of he error varance no beng consan over all cases s called heeroscedascy n conras o he condon of equal error varances, called homoscedascy. Heeroscedascy s nheren when he response n regresson analyss follows a dsrbuon n whch he varance s funconally relaed o he mean (so s relaed o a leas one predcor varable). The remedy for heeroscedascy s complcaed and, somemes, may no worh he effors. Transformaons of Y could ge you no roubles and Weghed Leas Squares s less ofen used because s hard o mplemen

66 THE ISSUE OF CORRELATED ERRORS

67 AUTOCORRELATION The basc mulple regresson models have assume ha he random error erms are ndependen normal random varables or, a leas, uncorrelaed random varables. In some felds for example n economcs, regresson applcaons may nvolve me seres ; he assumpon of uncorrelaed or ndependen error erms may no be approprae. In me seres daa, error erms are ofen (posvely) correlaed over me auo-correlaed or serally correlaed. Y β β x β x β x 0 k k ε

68 PROBLEMS OF AUTOCORRELATION Leas squares esmaes of regresson coeffcens are sll unbased bu no longer have mnmum varance MSE serously under-esmae varance of error erms Sandard errors of esmaed regresson coeffcens may serously under-esmae he rue sandard devaons of he esmaed regresson coeffcens; confden nervals of regresson coeffcens and of response means, herefore, may no have he correc coverage. and F ess may no longer applcable, have wrong sze.

69 FIRST-ORDER AUTOREGRESSIVE ERROR MODEL ),σ N( u where u x x x Y k k 0 0 's are ndependen : < ρ ρε ε ε β β β β

70 ),σ N( u where u x x x Y k k 0 0 's are ndependen : < ρ ρε ε ε β β β β Noe ha each error erm consss of a fracon of he prevous error erm plus a new dsurbance erm u ; he parameer ρ ofen posve - s called he auocorrelaon parameer.

71 Frs, we can easly expand from he defnon of he frs-order auoregressve error model o show ha each error erm s a lnear combnaon of curren and precedng dsurbance erms. Ths s used o prove ha he mean s zero and he varance s consan he varance s larger. 0 s s s 0 s s ρ σ ρ σ ) (ε σ 0 ) E(u ρ ) E(ε u ρε ε s s s u u u u u u u u u u u ) ( ) ( ρ ρ ρ ε ρ ρ ρε ρ ρ ε ρ ρε ρ

72 The error erms of he frs-order auoregressve model sll have mean zero and consan varance bu a posve covarance beween consecuve erms: ρ σ ρ },ε σ{ε ) ( 0 ) ( ρ σ ε σ ε E ),σ N( u where u x x x Y k k 0 0 's are ndependen : < ρ ρε ε ε β β β β

73 The auocorrelaon parameer s also he coeffcen of correlaon beween wo consecuve error erms: ρ ρ σ ρ σ ρ σ ρ ε σ ε σ ε ε σ ) ( ) ( ), (

74 s s s s ρ ρ σ ρ σ ρ σ ρ ε σ ε σ ε ε σ ) ( ) ( ), ( The coeffcen of correlaon below, of wo error erms ha are s perods apar, shows ha he error erms are also posvely correlaed bu he furher apar hey are he less he correlaon beween hem:

75 The ex presens a small smulaed daa se, pages , he regresson lne by leas squares mehod vares from case o case dependng on he value of he frs error erm. The esmaed slope ranges from a negave value o a posve value. I could be a serous problem unless we know how o deal wh auocorrelaon. And one could fnd example of me seres daa no only n economcs and busness bu also n bomedcal longudnal daa as well.

76 DURBIN-WATSON TEST The Durbn-Wason es for auocorrelaon assumes he frs-order auoregressve error model (wh values of predcor varables fxed) The es consss of deermnng wheher or no he auocorrelaon parameer (whch s also he coeffcen of correlaon beween consecuve error erms) s zero (f so, errors erms are equal o dsance erms whch are..d. normal): H 0 H A : : ρ ρ > 0 0

77 The Durbn-Wason es sasc D s obaned by frs usng ordnary leas squares mehod o f he regresson funcon, calculang resduals and hen D. Small values of D suppor he concluson of auocorrelaon because, under he frs-order auoregressve error model, adjacen error erms end o be of he same magnude (because hey are posvely correlaed) leadng o small dfferences and a small oal of hose dfference: D n ( e n e e )

78 Exac crcal values for he Durbn-Wason es sascs D are dffcul o oban, bu Durbn and Wason have provded lower and upper bounds d L and d U such ha an observed value of he es sasc D ousde hese bounds leads o a defnve decson. Values of he bound depend on he alpha level, number of predcor varables, and sample sze n (Table B7, page 675). If d If If L D > d D < d U L < D < d : Daa suppor H : Daa suppor H U : Tes s 0 A nconclus ve

79 SAS mplemenaon s very smple: Use opon DW PROC REG; MODEL Y X X X3/DW; An esmae of he auocorrelaon parameer s provded usng he followng formula: r n n e e e

80 When he presence of auocorrelaon s confrmed, he problem could be remeded by addng n anoher predcor varable or varables: one of he major cause s he omsson from he model of one or more key predcors:

81 BASIC SAS OPTIONS CORR: beween all varables n model saemen P: (or PRED) predced values R: (or RESID) resduals COVB: ( B for esmaed regresson coeffcen) CORRB: Varance-covarance marx (among B) CLB: confdence nerval for B CLM: confdence nerval for Mean Response CLI: confdence nerval for ndvdual predced.

82 SAS OPTIONS FOR SELECTION RSQUARE ADJRSQ MAXR AIC FORWARD BACKWARD STEPWISE

83 SAS OPTIONS FOR DIAGNOSTICS CORR: (ndvdual r >.7, say) STUDENT: sudenzed resduals PRESS: deleed resduals (square & add o oban he PRESS sasc) RSTUDENT: sudenzed deleed resduals ((-α/;n-p-) H: Leverages (p/n) COOKD: Cook s dsance (50% percenle of F(p,n-p) DFFITS: ( or p/n) DFBETAS: ( or / n)

84 Readngs & Exercses Readngs: A horough readng of he ex s secons (pp ) and skm over he example n secon 0.6 s recommended. Exercses: The followng exercses are good for pracce, all from chaper 0: 0.7, 0., 0.7, and 0..

85 Due As Homework #9. Refer o daase Cgarees, le Y log(nnal) and consder a model wh hree ndependen varables, X CPD, X Age, and X 3 Gender: a) F he mulple regresson model wh 3 predcor varables. Does appear ha all predcors should be reaned? b) Prepare an added-varable plo for X and X. Do your plos sugges ha he correspondng regresson relaonshps n (a) are approprae? Explan your answers. c) Oban he varance nflaon facors. Are here ndcaon ha serous mulcollneary problems exs here? Why or why no? d) Oban he resduals and prepare a normal probably plo; Also plo he resduals agans X and X, and draw conclusons. #9. Answer he 4 quesons of Exercse 9. usng daase Infans wh Y Brh Wegh, X Gesaonal Weeks, X Moher s Age, and X 3 Toxema (oxema 0/ s a pregnancy condon resulng from meabolc dsorder).

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6)

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6) Econ7 Appled Economercs Topc 5: Specfcaon: Choosng Independen Varables (Sudenmund, Chaper 6 Specfcaon errors ha we wll deal wh: wrong ndependen varable; wrong funconal form. Ths lecure deals wh wrong ndependen

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction ECOOMICS 35* -- OTE 9 ECO 35* -- OTE 9 F-Tess and Analyss of Varance (AOVA n he Smple Lnear Regresson Model Inroducon The smple lnear regresson model s gven by he followng populaon regresson equaon, or

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

January Examinations 2012

January Examinations 2012 Page of 5 EC79 January Examnaons No. of Pages: 5 No. of Quesons: 8 Subjec ECONOMICS (POSTGRADUATE) Tle of Paper EC79 QUANTITATIVE METHODS FOR BUSINESS AND FINANCE Tme Allowed Two Hours ( hours) Insrucons

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

Panel Data Regression Models

Panel Data Regression Models Panel Daa Regresson Models Wha s Panel Daa? () Mulple dmensoned Dmensons, e.g., cross-secon and me node-o-node (c) Pongsa Pornchawseskul, Faculy of Economcs, Chulalongkorn Unversy (c) Pongsa Pornchawseskul,

More information

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA Mchaela Chocholaá Unversy of Economcs Braslava, Slovaka Inroducon (1) one of he characersc feaures of sock reurns

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that THEORETICAL AUTOCORRELATIONS Cov( y, y ) E( y E( y))( y E( y)) ρ = = Var( y) E( y E( y)) =,, L ρ = and Cov( y, y ) s ofen denoed by whle Var( y ) f ofen denoed by γ. Noe ha γ = γ and ρ = ρ and because

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

Lecture 6: Learning for Control (Generalised Linear Regression)

Lecture 6: Learning for Control (Generalised Linear Regression) Lecure 6: Learnng for Conrol (Generalsed Lnear Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure 6: RLSC - Prof. Sehu Vjayakumar Lnear Regresson

More information

Advanced time-series analysis (University of Lund, Economic History Department)

Advanced time-series analysis (University of Lund, Economic History Department) Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

Lecture VI Regression

Lecture VI Regression Lecure VI Regresson (Lnear Mehods for Regresson) Conens: Lnear Mehods for Regresson Leas Squares, Gauss Markov heorem Recursve Leas Squares Lecure VI: MLSC - Dr. Sehu Vjayakumar Lnear Regresson Model M

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations Chaper 6: Ordnary Leas Squares Esmaon Procedure he Properes Chaper 6 Oulne Cln s Assgnmen: Assess he Effec of Sudyng on Quz Scores Revew o Regresson Model o Ordnary Leas Squares () Esmaon Procedure o he

More information

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10)

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10) Publc Affars 974 Menze D. Chnn Fall 2009 Socal Scences 7418 Unversy of Wsconsn-Madson Problem Se 2 Answers Due n lecure on Thursday, November 12. " Box n" your answers o he algebrac quesons. 1. Consder

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

Time-interval analysis of β decay. V. Horvat and J. C. Hardy

Time-interval analysis of β decay. V. Horvat and J. C. Hardy Tme-nerval analyss of β decay V. Horva and J. C. Hardy Work on he even analyss of β decay [1] connued and resuled n he developmen of a novel mehod of bea-decay me-nerval analyss ha produces hghly accurae

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015)

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015) 5h Inernaonal onference on Advanced Desgn and Manufacurng Engneerng (IADME 5 The Falure Rae Expermenal Sudy of Specal N Machne Tool hunshan He, a, *, La Pan,b and Bng Hu 3,c,,3 ollege of Mechancal and

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth Should Exac Index umbers have Sandard Errors? Theory and Applcaon o Asan Growh Rober C. Feensra Marshall B. Rensdorf ovember 003 Proof of Proposon APPEDIX () Frs, we wll derve he convenonal Sao-Vara prce

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Machine Learning Linear Regression

Machine Learning Linear Regression Machne Learnng Lnear Regresson Lesson 3 Lnear Regresson Bascs of Regresson Leas Squares esmaon Polynomal Regresson Bass funcons Regresson model Regularzed Regresson Sascal Regresson Mamum Lkelhood (ML)

More information

Chapter 5. The linear fixed-effects estimators: matrix creation

Chapter 5. The linear fixed-effects estimators: matrix creation haper 5 he lnear fed-effecs esmaors: mar creaon In hs chaper hree basc models and he daa marces needed o creae esmaors for hem are defned. he frs s ermed he cross-secon model: alhough ncorporaes some panel

More information

CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING

CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING 4. Inroducon The repeaed measures sudy s a very commonly used expermenal desgn n oxcy esng because no only allows one o nvesgae he effecs of he oxcans,

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas)

Lecture 18: The Laplace Transform (See Sections and 14.7 in Boas) Lecure 8: The Lalace Transform (See Secons 88- and 47 n Boas) Recall ha our bg-cure goal s he analyss of he dfferenal equaon, ax bx cx F, where we emloy varous exansons for he drvng funcon F deendng on

More information

Analysis And Evaluation of Econometric Time Series Models: Dynamic Transfer Function Approach

Analysis And Evaluation of Econometric Time Series Models: Dynamic Transfer Function Approach 1 Appeared n Proceedng of he 62 h Annual Sesson of he SLAAS (2006) pp 96. Analyss And Evaluaon of Economerc Tme Seres Models: Dynamc Transfer Funcon Approach T.M.J.A.COORAY Deparmen of Mahemacs Unversy

More information

CHAPTER 5: MULTIVARIATE METHODS

CHAPTER 5: MULTIVARIATE METHODS CHAPER 5: MULIVARIAE MEHODS Mulvarae Daa 3 Mulple measuremens (sensors) npus/feaures/arbues: -varae N nsances/observaons/eamples Each row s an eample Each column represens a feaure X a b correspons o he

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

FI 3103 Quantum Physics

FI 3103 Quantum Physics /9/4 FI 33 Quanum Physcs Aleander A. Iskandar Physcs of Magnesm and Phooncs Research Grou Insu Teknolog Bandung Basc Conces n Quanum Physcs Probably and Eecaon Value Hesenberg Uncerany Prncle Wave Funcon

More information

2. SPATIALLY LAGGED DEPENDENT VARIABLES

2. SPATIALLY LAGGED DEPENDENT VARIABLES 2. SPATIALLY LAGGED DEPENDENT VARIABLES In hs chaper, we descrbe a sascal model ha ncorporaes spaal dependence explcly by addng a spaally lagged dependen varable y on he rgh-hand sde of he regresson equaon.

More information

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management P age NPTEL Proec Economerc Modellng Vnod Gua School of Managemen Module23: Granger Causaly Tes Lecure35: Granger Causaly Tes Rudra P. Pradhan Vnod Gua School of Managemen Indan Insue of Technology Kharagur,

More information

Comparison of Supervised & Unsupervised Learning in βs Estimation between Stocks and the S&P500

Comparison of Supervised & Unsupervised Learning in βs Estimation between Stocks and the S&P500 Comparson of Supervsed & Unsupervsed Learnng n βs Esmaon beween Socks and he S&P500 J. We, Y. Hassd, J. Edery, A. Becker, Sanford Unversy T I. INTRODUCTION HE goal of our proec s o analyze he relaonshps

More information

Bernoulli process with 282 ky periodicity is detected in the R-N reversals of the earth s magnetic field

Bernoulli process with 282 ky periodicity is detected in the R-N reversals of the earth s magnetic field Submed o: Suden Essay Awards n Magnecs Bernoull process wh 8 ky perodcy s deeced n he R-N reversals of he earh s magnec feld Jozsef Gara Deparmen of Earh Scences Florda Inernaonal Unversy Unversy Park,

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Kayode Ayinde Department of Pure and Applied Mathematics, Ladoke Akintola University of Technology P. M. B. 4000, Ogbomoso, Oyo State, Nigeria

Kayode Ayinde Department of Pure and Applied Mathematics, Ladoke Akintola University of Technology P. M. B. 4000, Ogbomoso, Oyo State, Nigeria Journal of Mahemacs and Sascs 3 (4): 96-, 7 ISSN 549-3644 7 Scence Publcaons A Comparave Sudy of he Performances of he OLS and some GLS Esmaors when Sochasc egressors are boh Collnear and Correlaed wh

More information

Data Collection Definitions of Variables - Conceptualize vs Operationalize Sample Selection Criteria Source of Data Consistency of Data

Data Collection Definitions of Variables - Conceptualize vs Operationalize Sample Selection Criteria Source of Data Consistency of Data Apply Sascs and Economercs n Fnancal Research Obj. of Sudy & Hypoheses Tesng From framework objecves of sudy are needed o clarfy, hen, n research mehodology he hypoheses esng are saed, ncludng esng mehods.

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

PhD/MA Econometrics Examination. January, 2019

PhD/MA Econometrics Examination. January, 2019 Economercs Comprehensve Exam January 2019 Toal Tme: 8 hours MA sudens are requred o answer from A and B. PhD/MA Economercs Examnaon January, 2019 PhD sudens are requred o answer from A, B, and C. The answers

More information

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance INF 43 3.. Repeon Anne Solberg (anne@f.uo.no Bayes rule for a classfcaon problem Suppose we have J, =,...J classes. s he class label for a pxel, and x s he observed feaure vecor. We can use Bayes rule

More information

Let s treat the problem of the response of a system to an applied external force. Again,

Let s treat the problem of the response of a system to an applied external force. Again, Page 33 QUANTUM LNEAR RESPONSE FUNCTON Le s rea he problem of he response of a sysem o an appled exernal force. Agan, H() H f () A H + V () Exernal agen acng on nernal varable Hamlonan for equlbrum sysem

More information

Chapter Lagrangian Interpolation

Chapter Lagrangian Interpolation Chaper 5.4 agrangan Inerpolaon Afer readng hs chaper you should be able o:. dere agrangan mehod of nerpolaon. sole problems usng agrangan mehod of nerpolaon and. use agrangan nerpolans o fnd deraes and

More information

Chapter 8 Dynamic Models

Chapter 8 Dynamic Models Chaper 8 Dnamc odels 8. Inroducon 8. Seral correlaon models 8.3 Cross-seconal correlaons and me-seres crosssecon models 8.4 me-varng coeffcens 8.5 Kalman fler approach 8. Inroducon When s mporan o consder

More information

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

Computing Relevance, Similarity: The Vector Space Model

Computing Relevance, Similarity: The Vector Space Model Compung Relevance, Smlary: The Vecor Space Model Based on Larson and Hears s sldes a UC-Bereley hp://.sms.bereley.edu/courses/s0/f00/ aabase Managemen Sysems, R. Ramarshnan ocumen Vecors v ocumens are

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC

CH.3. COMPATIBILITY EQUATIONS. Continuum Mechanics Course (MMC) - ETSECCPB - UPC CH.3. COMPATIBILITY EQUATIONS Connuum Mechancs Course (MMC) - ETSECCPB - UPC Overvew Compably Condons Compably Equaons of a Poenal Vecor Feld Compably Condons for Infnesmal Srans Inegraon of he Infnesmal

More information

ABSTRACT KEYWORDS. Bonus-malus systems, frequency component, severity component. 1. INTRODUCTION

ABSTRACT KEYWORDS. Bonus-malus systems, frequency component, severity component. 1. INTRODUCTION EERAIED BU-MAU YTEM ITH A FREQUECY AD A EVERITY CMET A IDIVIDUA BAI I AUTMBIE IURACE* BY RAHIM MAHMUDVAD AD HEI HAAI ABTRACT Frangos and Vronos (2001) proposed an opmal bonus-malus sysems wh a frequency

More information

EDO UNIVERSITY, IYAMHO EDO STATE, NIGERIA DEPARTMENT OF ECONOMICS

EDO UNIVERSITY, IYAMHO EDO STATE, NIGERIA DEPARTMENT OF ECONOMICS EDO UNIVERSITY, IYAMHO EDO STATE, NIGERIA DEPARTMENT OF ECONOMICS ECO 4: APPLIED ECONOMETRICS INSTRUCTOR: LECTURES: OFFICE HOURS: Davd Umoru, Emal: davd.umoru@edounversy.edu.ng Alernave emal: davd.umoru@yahoo.com

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

Endogeneity. Is the term given to the situation when one or more of the regressors in the model are correlated with the error term such that

Endogeneity. Is the term given to the situation when one or more of the regressors in the model are correlated with the error term such that s row Endogeney Is he erm gven o he suaon when one or more of he regressors n he model are correlaed wh he error erm such ha E( u 0 The 3 man causes of endogeney are: Measuremen error n he rgh hand sde

More information

Robustness of DEWMA versus EWMA Control Charts to Non-Normal Processes

Robustness of DEWMA versus EWMA Control Charts to Non-Normal Processes Journal of Modern Appled Sascal Mehods Volume Issue Arcle 8 5--3 Robusness of D versus Conrol Chars o Non- Processes Saad Saeed Alkahan Performance Measuremen Cener of Governmen Agences, Insue of Publc

More information

Machine Learning 2nd Edition

Machine Learning 2nd Edition INTRODUCTION TO Lecure Sldes for Machne Learnng nd Edon ETHEM ALPAYDIN, modfed by Leonardo Bobadlla and some pars from hp://www.cs.au.ac.l/~aparzn/machnelearnng/ The MIT Press, 00 alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/mle

More information

Chapter 6: AC Circuits

Chapter 6: AC Circuits Chaper 6: AC Crcus Chaper 6: Oulne Phasors and he AC Seady Sae AC Crcus A sable, lnear crcu operang n he seady sae wh snusodal excaon (.e., snusodal seady sae. Complee response forced response naural response.

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Time Scale Evaluation of Economic Forecasts

Time Scale Evaluation of Economic Forecasts CENTRAL BANK OF CYPRUS EUROSYSTEM WORKING PAPER SERIES Tme Scale Evaluaon of Economc Forecass Anons Mchs February 2014 Worng Paper 2014-01 Cenral Ban of Cyprus Worng Papers presen wor n progress by cenral

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

A First Guide to Hypothesis Testing in Linear Regression Models. A Generic Linear Regression Model: Scalar Formulation

A First Guide to Hypothesis Testing in Linear Regression Models. A Generic Linear Regression Model: Scalar Formulation ECON 5* -- A rs Gude o Hypoess Tesng MG Abbo A rs Gude o Hypoess Tesng n Lnear Regresson Models A Generc Lnear Regresson Model: Scalar mulaon e - populaon ( sample observaon, e scalar fmulaon of e PRE

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

On computing differential transform of nonlinear non-autonomous functions and its applications

On computing differential transform of nonlinear non-autonomous functions and its applications On compung dfferenal ransform of nonlnear non-auonomous funcons and s applcaons Essam. R. El-Zahar, and Abdelhalm Ebad Deparmen of Mahemacs, Faculy of Scences and Humanes, Prnce Saam Bn Abdulazz Unversy,

More information

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar

Anomaly Detection. Lecture Notes for Chapter 9. Introduction to Data Mining, 2 nd Edition by Tan, Steinbach, Karpatne, Kumar Anomaly eecon Lecure Noes for Chaper 9 Inroducon o aa Mnng, 2 nd Edon by Tan, Senbach, Karpane, Kumar 2/14/18 Inroducon o aa Mnng, 2nd Edon 1 Anomaly/Ouler eecon Wha are anomales/oulers? The se of daa

More information

Methods for the estimation of missing values in time series

Methods for the estimation of missing values in time series Edh Cowan Unversy Research Onlne Theses: Docoraes and Masers Theses 6 Mehods for he esmaon of mssng values n me seres Davd S. Fung Edh Cowan Unversy Recommended Caon Fung, D. S. (6). Mehods for he esmaon

More information

The Performance of Optimum Response Surface Methodology Based on MM-Estimator

The Performance of Optimum Response Surface Methodology Based on MM-Estimator The Performance of Opmum Response Surface Mehodology Based on MM-Esmaor Habshah Md, Mohd Shafe Musafa, Anwar Frano Absrac The Ordnary Leas Squares (OLS) mehod s ofen used o esmae he parameers of a second-order

More information

Polymerization Technology Laboratory Course

Polymerization Technology Laboratory Course Prakkum Polymer Scence/Polymersaonsechnk Versuch Resdence Tme Dsrbuon Polymerzaon Technology Laboraory Course Resdence Tme Dsrbuon of Chemcal Reacors If molecules or elemens of a flud are akng dfferen

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

Normal Random Variable and its discriminant functions

Normal Random Variable and its discriminant functions Noral Rando Varable and s dscrnan funcons Oulne Noral Rando Varable Properes Dscrnan funcons Why Noral Rando Varables? Analycally racable Works well when observaon coes for a corruped snle prooype 3 The

More information

1 Constant Real Rate C 1

1 Constant Real Rate C 1 Consan Real Rae. Real Rae of Inees Suppose you ae equally happy wh uns of he consumpon good oday o 5 uns of he consumpon good n peod s me. C 5 Tha means you ll be pepaed o gve up uns oday n eun fo 5 uns

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

MODELING TIME-VARYING TRADING-DAY EFFECTS IN MONTHLY TIME SERIES

MODELING TIME-VARYING TRADING-DAY EFFECTS IN MONTHLY TIME SERIES MODELING TIME-VARYING TRADING-DAY EFFECTS IN MONTHLY TIME SERIES Wllam R. Bell, Census Bureau and Donald E. K. Marn, Howard Unversy and Census Bureau Donald E. K. Marn, Howard Unversy, Washngon DC 0059

More information

Dynamic Team Decision Theory

Dynamic Team Decision Theory Dynamc Team Decson Theory EECS 558 Proec Repor Shruvandana Sharma and Davd Shuman December, 005 I. Inroducon Whle he sochasc conrol problem feaures one decson maker acng over me, many complex conrolled

More information

II. Light is a Ray (Geometrical Optics)

II. Light is a Ray (Geometrical Optics) II Lgh s a Ray (Geomercal Opcs) IIB Reflecon and Refracon Hero s Prncple of Leas Dsance Law of Reflecon Hero of Aleandra, who lved n he 2 nd cenury BC, posulaed he followng prncple: Prncple of Leas Dsance:

More information

Economics 120C Final Examination Spring Quarter June 11 th, 2009 Version A

Economics 120C Final Examination Spring Quarter June 11 th, 2009 Version A Suden Name: Economcs 0C Sprng 009 Suden ID: Name of Suden o your rgh: Name of Suden o your lef: Insrucons: Economcs 0C Fnal Examnaon Sprng Quarer June h, 009 Verson A a. You have 3 hours o fnsh your exam.

More information

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer

12d Model. Civil and Surveying Software. Drainage Analysis Module Detention/Retention Basins. Owen Thornton BE (Mech), 12d Model Programmer d Model Cvl and Surveyng Soware Dranage Analyss Module Deenon/Reenon Basns Owen Thornon BE (Mech), d Model Programmer owen.hornon@d.com 4 January 007 Revsed: 04 Aprl 007 9 February 008 (8Cp) Ths documen

More information

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective Forecasng cusomer behavour n a mul-servce fnancal organsaon: a profably perspecve A. Audzeyeva, Unversy of Leeds & Naonal Ausrala Group Europe, UK B. Summers, Unversy of Leeds, UK K.R. Schenk-Hoppé, Unversy

More information

How about the more general "linear" scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )?

How about the more general linear scalar functions of scalars (i.e., a 1st degree polynomial of the following form with a constant term )? lmcd Lnear ransformaon of a vecor he deas presened here are que general hey go beyond he radonal mar-vecor ype seen n lnear algebra Furhermore, hey do no deal wh bass and are equally vald for any se of

More information

Additive Outliers (AO) and Innovative Outliers (IO) in GARCH (1, 1) Processes

Additive Outliers (AO) and Innovative Outliers (IO) in GARCH (1, 1) Processes Addve Oulers (AO) and Innovave Oulers (IO) n GARCH (, ) Processes MOHAMMAD SAID ZAINOL, SITI MERIAM ZAHARI, KAMARULZAMMAN IBRAHIM AZAMI ZAHARIM, K. SOPIAN Cener of Sudes for Decson Scences, FSKM, Unvers

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

High frequency analysis of lead-lag relationships between financial markets de Jong, Frank; Nijman, Theo

High frequency analysis of lead-lag relationships between financial markets de Jong, Frank; Nijman, Theo Tlburg Unversy Hgh frequency analyss of lead-lag relaonshps beween fnancal markes de Jong, Frank; Nman, Theo Publcaon dae: 1995 Lnk o publcaon Caon for publshed verson (APA): de Jong, F. C. J. M., & Nman,

More information