Probabilistic Reasoning; Graphical models

Size: px
Start display at page:

Download "Probabilistic Reasoning; Graphical models"

Transcription

1 Ovrvw. robablsc Rasonng/Graphcal modls. Imporanc Samplng 3. Markov Chan Mon Carlo: Gbbs Samplng 4. Samplng n prsnc of Drmnsm 5. Rao-Blackwllsaon 6. AND/OR mporanc samplng

2 Ovrvw. robablsc Rasonng/Graphcal modls. Imporanc Samplng 3. Markov Chan Mon Carlo: Gbbs Samplng 4. Samplng n prsnc of Drmnsm 5. Cus-basd Varanc Rducon 6. AND/OR mporanc samplng

3 robablsc Rasonng; Graphcal modls Graphcal modls: Baysan nwork consran nworks md nwork Qurs Eac algorhm usng nfrnc sarch and hybrds Graph paramrs: r-wdh cycl-cus w-cus

4 Qurs robably of vdnc or paron funcon var pa osror margnal blfs: Mos robabl Eplanaon n var j n n var j Z j j pa pa j j C * argma

5 Appromaon Snc nfrnc sarch and hybrds ar oo pnsv whn graph s dns; hgh rwdh hn: Boundng nfrnc: mn-buck and mn-clusrng Blf propagaon Boundng sarch: Samplng Goal: an anym schm 5

6 Ovrvw. robablsc Rasonng/Graphcal modls. Imporanc Samplng 3. Markov Chan Mon Carlo: Gbbs Samplng 4. Samplng n prsnc of Drmnsm 5. Rao-Blackwllsaon 6. AND/OR mporanc samplng

7 Ouln Dfnons and Background on Sascs Thory of mporanc samplng Lklhood wghng Sa-of-h-ar mporanc samplng chnqus 7

8 A sampl Gvn a s of varabls ={... n } a sampl dnod by S s an nsanaon of all varabls: S... n 8

9 How o draw a sampl? Unvara dsrbuon Eampl: Gvn random varabl havng doman {0 } and a dsrbuon = Task: Gnra sampls of from. How? draw random numbr r [0 ] If r < 0.3 hn s =0 Els s = 9

10 How o draw a sampl? Mul-vara dsrbuon L ={.. n } b a s of varabls Eprss h dsrbuon n produc form... n... n Sampl varabls on by on from lf o rgh along h ordrng dcad by h produc form. Baysan nwork lraur: Logc samplng 0

11 Samplng for rob. Infrnc Ouln Logc Samplng Imporanc Samplng Lklhood Samplng Choosng a roposal Dsrbuon Markov Chan Mon Carlo MCMC Mropols-Hasngs Gbbs samplng Varanc Rducon

12 Logc Samplng: No Evdnc Hnron 988 Inpu: Baysan nwork = { N } N- #nods T - # sampls Oupu: T sampls rocss nods n opologcal ordr frs procss h ancsors of a nod hn h nod slf:. For = 0 o T. For = 0 o N 3. sampl from pa

13 Logc samplng ampl No Evdnc // gnra sampl k.sampl.sampl 3.Sampl 4.Sampl 3 4 from from from 3 from

14 Logc Samplng w/ Evdnc Inpu: Baysan nwork = { N } N- #nods E vdnc T - # sampls Oupu: T sampls conssn wh E. For = o T. For = o N 3. sampl from pa 4. If n E and rjc sampl: 5. Goo Sp. 4

15 Logc Samplng ampl Evdnc : // gnra sampl k.sampl.sampl 3.Sampl 4. If 3 and sar from ohrws 5.Sampl 3 from from from 0 rjc sampl 4 3 from 4 3 5

16 Epcd valu and Varanc Epcd valu: Gvn a probably dsrbuon and a funcon g dfnd ovr a s of varabls = { n } h pcd valu of g w.r.. s E [ g ] g Varanc: Th varanc of g w.r.. s: Var [ g ] g E [ g ] 6

17 Mon Carlo Esma Esmaor: An smaor s a funcon of h sampls. I producs an sma of h unknown paramr of h samplng dsrbuon. Gvn..d. sampls S gˆ T T S h M oncarlo sma of g S S E T drawn from [g] s gvn by : 7

18 Eampl: Mon Carlo sma Gvn: A dsrbuon = g = 40 f quals 0 = 50 f quals. Esma E [g]= =47. Gnra k sampls from : 0000 gˆ 40# sampls 0 50# # sampls sampls

19 Ouln Dfnons and Background on Sascs Thory of mporanc samplng Lklhood wghng Sa-of-h-ar mporanc samplng chnqus 9

20 Imporanc samplng: Man da Eprss qury as h pcd valu of a random varabl w.r.. o a dsrbuon Q. Gnra random sampls from Q. Esma h pcd valu from h gnrad sampls usng a mon carlo smaor avrag. 0

21 Imporanc samplng for ˆ : ] [ \ Z Q z w T z w E z Q z E z Q z Q z z z Q z E Z L T Q Q z z whr z M oncarlo sma as : w can rwr Thn 0 0 sasfyng proposaldsrbuon L QZb a

22 roprs of IS sma of Convrgnc: by law of larg numbrs Unbasd. Varanc: ˆ T a. s w z. for T T Var E Q Q [ ˆ ] ˆ Var Q T N w z Var Q [ w z] T

23 roprs of IS sma of Man Squard Error of h smaor MSE Q ˆ E ˆ Q E ˆ [ ] Var ˆ Q Q ˆ Var Var Q Q [ w ] T Ths quany nclosd n h bracks s zro bcaus h pcd valu of h smaor quals h pcd valu of g

24 Esmang 4 E basd : Esma s ˆ ˆ Rao sma: IS. Ida : Esma numraor and dnomnaor by and 0 ohrws. z conans whch s f drac - dla funcon z b a L T k T k wz wz z z Q z E z Q z z E z z z k k k Q Q z z

25 roprs of h IS smaor for Convrgnc: By Wak law of larg numbrs as T Asympocally unbasd Varanc lm T E Hardr o analyz [ ] Lu suggss a masur calld Effcv sampl sz 5

26 Gnrang sampls from Q No rsrcons on how o Typcally prss Q n produc form: QZ=QZ QZ Z.QZ n Z..Z n- Sampl along h ordr Z..Z n Eampl: Z QZ =0.0.8 Z QZ Z = Z 3 QZ 3 Z Z =QZ 3 =0.50.5

27 Ouln Dfnons and Background on Sascs Thory of mporanc samplng Lklhood wghng Sa-of-h-ar mporanc samplng chnqus 7

28 Lklhood Wghng Fung and Chang 990; Shachr and o 990 Is an nsanc of mporanc samplng! Clampng vdnc+ logc samplng+ wghng sampls by vdnc lklhood Works wll for lkly vdnc! 8

29 Lklhood Wghng: Samplng Sampl n opologcal ordr ovr! Clamp vdnc Sampl pa pa s a look-up n CT! 9

30 Lklhood Wghng: roposal Dsrbuon 30 E E j j E E E E j j n E j j pa pa pa pa Q w Wghs pa E Q.. : : \ \ \ \ Gvn a sampl Q. Evdnc and Baysan nwork: Gvn a : Eampl Noc: Q s anohr Baysan nwork

31 Lklhood Wghng: Esmas 3 T w T ˆ Esma : zro ohrws and quals f ˆ ˆ ˆ T T g w g w Esma osror Margnals:

32 Lklhood Wghng Convrgs o ac posror margnals Gnras Sampls Fas Samplng dsrbuon s clos o pror spcally f E Laf Nods Incrasng samplng varanc Convrgnc may b slow Many sampls wh =0 rjcd 3

33 Ouln Dfnons and Background on Sascs Thory of mporanc samplng Lklhood wghng Error smaon Sa-of-h-ar mporanc samplng chnqus 33

34

35

36

37 absolu

38 Ouln Dfnons and Background on Sascs Thory of mporanc samplng Lklhood wghng Sa-of-h-ar mporanc samplng chnqus 38

39 roposal slcon On should ry o slc a proposal ha s as clos as possbl o h posror dsrbuon. smaor zro- varanc o hav a 0 ] [ ˆ z z Q z Q z z Q z z Q z Q z N T z w Var Var Z z Q Q

40 rfc samplng usng Buck Algorhm: Elmnaon Run Buck lmnaon on h problm along an ordrng o= N... Sampl along h rvrs ordrng:.. N A ach varabl rcovr h probably... - by rfrrng o h buck.

41 4 Buck Elmnaon 0 0 a a d b c c b b a d a c a b a a 0 0 d c b b a d c b a b a c a 0 Elmnaon Ordr: dbc Qury: D: E: B: C: A: d D b a d b a f b a d c b 0 c b c b f E b E D B c b f b a f a b c a f 0 a f A p a C a a c c B C c a f a c a f a b DAB EBC BAC CA A b a f D c b f E c a f B f C a A D E C B Buck Tr D E B C A Orgnal Funcons Mssags Tm and spac pw*

42 buck B: buck C: buck D: buck E: buck A: Buck lmnaon BE Algorhm lm-bl Dchr 996 CA a b BA DBA BC h B ADC h C AD h D A h E a Elmnaon opraor S 4 B C D E A

43 Samplng from h oupu of BE Dchr 00 buck B: S A a D dc c n h buck Sampl : B b QB a d B a d BA DBA BC B a b c buck C: CA h B ADC S A a D d n h buck Sampl :C c QC a d C A h B a d C buck D: h C AD S A a n h buck Sampl:D d QD a h C ad buck E: h D A Evdnc buck :gnor buck A: A h E A QA A h E A Sampl : A a QA S 43

44 Mn-bucks: local nfrnc Compuaon n a buck s m and spac ponnal n h numbr of varabls nvolvd Thrfor paron funcons n a buck no mn-bucks on smallr numbr of varabls Can conrol h sz of ach mn-buck yldng polynomal comply. S 44

45 Mn-Buck Elmnaon buck B: Σ B BC Mn-bucks Σ B BA DBA Spac and Tm consrans: Mamum scop sz of h nw funcon gnrad should b boundd by buck C: buck D: CA h B C h B AD BE gnras a funcon havng scop sz 3. So canno b usd. buck E: h C A buck A: A h E A h D A Appromaon of S 45 45

46 Samplng from h oupu of MBE buck B: BC BA DBA buck C: CA h B C buck D: h B AD buck E: buck A: h E A h C A h D A Samplng s sam as n BE-samplng cp ha now w consruc Q from a randomly slcd mnbuck S 46 46

47 IJG-Samplng Goga and Dchr 005 Irav Jon Graph ropagaon IJG A Gnralzd Blf ropagaon schm Ydda al. 00 IJG ylds br appromaons of E han MBE Dchr Kask and Mascu 00 Oupu of IJG s sam as mn-buck clusrs Currnly h bs prformng IS schm!

48 Currn Rsarch quson Gvn a Baysan nwork wh vdnc or a Markov nwork rprsnng funcon gnra anohr Baysan nwork rprsnng a funcon Q from a famly of dsrbuons rsrcd by srucur such ha Q s closs o. Currn approachs Mn-bucks Ijgp Boh Eprmnd bu nd o b jusfd horcally.

49 Algorhm: Approma Samplng Run IJG or MBE A ach branch pon compu h dg probabls by consulng oupu of IJG or MBE Rjcon roblm: Som assgnmns gnrad ar non soluons

50 k ˆ R ' UpdaQ N ˆ ˆE Q... z Gnra sampls z o k do For 0 ˆ... Q roposal Inal k N E urn End Q Q k Q z w E from E Z pa Z Q Z pa Z Q Z Q Z k k N j k k n n Adapv Imporanc Samplng

51 Adapv Imporanc Samplng Gnral cas Gvn k proposal dsrbuons Tak N sampls ou of ach dsrbuon Approma ˆ k k j Avg wgh jh proposal

52 Esmang Q'z ' Q Z Q' Z whr ach Q'Z Q' Z Z.. Z pa Z s smad by mporanc samplng -... Q' Z pa n Z n

53 Ovrvw. robablsc Rasonng/Graphcal modls. Imporanc Samplng 3. Markov Chan Mon Carlo: Gbbs Samplng 4. Samplng n prsnc of Drmnsm 5. Rao-Blackwllsaon 6. AND/OR mporanc samplng

54 Markov Chan 3 4 A Markov chan s a dscr random procss wh h propry ha h n sa dpnds only on h currn sa Markov ropry:... If - dos no dpnd on m homognous and sa spac s fn hn s ofn prssd as a ranson funcon aka ranson mar 54

55 Eampl: Drunkard s Walk a random walk on h numbr ln whr a ach sp h poson may chang by + or wh qual probably 3 D {0...} n n 0.5 n 0.5 ranson mar 55

56 Eampl: Wahr Modl ran ran ran sun ran D { rany sunny} rany sunny rany sunny ranson mar 56

57 Mul-Varabl Sysm } D dscr { 3 sa s an assgnmn of valus o all h varabls fn {... n } 57

58 Baysan Nwork Sysm Baysan Nwork s a rprsnaon of h jon probably dsrbuon ovr or mor varabls { 3 } { } 3 58

59 Saonary Dsrbuon Esnc If h Markov chan s m-homognous hn h vcor s a saonary dsrbuon aka nvaran or qulbrum dsrbuon aka fd pon f s nrs sum up o and sasfy: D Fn sa spac Markov chan has a unqu saonary dsrbuon f and only f: Th chan s rrducbl All of s sas ar posv rcurrn j j 59

60 Irrducbl A sa s rrducbl f undr h ranson rul on has nonzro probably of movng from o any ohr sa and hn comng back n a fn numbr of sps If on sa s rrducbl hn all h sas mus b rrducbl Lu Ch. pp. 49 Df... 60

61 Rcurrn A sa s rcurrn f h chan rurns o wh probably L M b h pcd numbr of sps o rurn o sa Sa s posv rcurrn f M s fn Th rcurrn sas n a fn sa chan ar posv rcurrn. 6

62 Saonary Dsrbuon Convrgnc Consdr nfn Markov chan: n n 0 If h chan s boh rrducbl and aprodc hn: lm n n Inal sa s no mporan n h lm Th mos usful faur of a good Markov chan s s fas forgfulnss of s pas Lu Ch.. 0 n 6

63 Aprodc Dfn d = g.c.d.{n > 0 s possbl o go from o n n sps}. Hr g.c.d. mans h gras common dvsor of h ngrs n h s. If d= for hn chan s aprodc osv rcurrn aprodc sas ar rgodc 63

64 Markov Chan Mon Carlo How do w sma.g.? Gnra sampls ha form Markov Chan wh saonary dsrbuon = Esma from sampls obsrvd sas: vsd sas 0 n can b vwd as sampls from dsrbuon T T lm T 64

65 MCMC Summary Convrgnc s guarand n h lm Inal sa s no mporan bu ypcally w hrow away frs K sampls - burn-n Sampls ar dpndn no..d. Convrgnc mng ra may b slow Th srongr corrlaon bwn sas h slowr convrgnc! 65

66 Gbbs Samplng Gman&Gman984 Gbbs samplr s an algorhm o gnra a squnc of sampls from h jon probably dsrbuon of wo or mor random varabls Sampl nw varabl valu on varabl a a m from h varabl s condonal dsrbuon:..... } \ n Sampls form a Markov chan wh saonary dsrbuon 66

67 Gbbs Samplng: Illusraon Th procss of Gbbs samplng can b undrsood as a random walk n h spac of all nsanaons of = rmmbr drunkard s walk: In on sp w can rach nsanaons ha dffr from currn on by valu assgnmn o a mos on varabl assum randomzd choc of varabls.

68 Ordrd Gbbs Samplr Gnra sampl + from : In shor for = o N: 68 \ sampld from N N N N N N rocss All Varabls In Som Ordr

69 Transon robabls n BN Gvn Markov blank parns chldrn and hr parns s ndpndn of all ohr nods \ \ Markov blank: markov pa ch pa markov pa j ch Compuaon s lnar n h sz of Markov blank! : j pa j ch j j j 69

70 Ordrd Gbbs Samplng Algorhm arl988 Inpu: E= Oupu: T sampls { } F vdnc E= nalz 0 a random. For = o T compu sampls. For = o N loop hrough varabls 3. + markov 4. End For 5. End For

71 Gbbs Samplng Eampl - BN {... 9} E { 9} 3 6 = 0 6 = 6 0 = = = = = = 5 0 7

72 Gbbs Samplng Eampl - BN {... 9} E { 9}

73 Answrng Qurs =? Mhod : coun # of sampls whr = hsogram smaor: T markov T T T Drac dla f-n Mhod : avrag probably mur smaor: Mur smaor convrgs fasr consdr smas for h unobsrvd valus of ; prov va Rao-Blackwll horm

74 Rao-Blackwll Thorm Rao-Blackwll Thorm: L random varabl s b composd of wo groups of varabls R and L. Thn for h jon dsrbuon RL and funcon g h followng rsul appls Var[ E{ g R L} Var[ g R] for a funcon of nrs g.g. h man or covaranc Caslla&Robr996 Lu. al horm maks a wak proms bu works wll n pracc! mprovmn dpnds h choc of R and L 74

75 Imporanc vs. Gbbs T T T Q g T g Q g T g ˆ ˆ ˆ w Gbbs: Imporanc:

76 Gbbs Samplng: Convrgnc Sampl from ` Convrgs ff chan s rrducbl and rgodc Inuon - mus b abl o plor all sas: f and j ar srongly corrlad =0 j =0 hn w canno plor sas wh = and j = All condons ar sasfd whn all probabls ar posv Convrgnc ra can b characrzd by h scond gn-valu of ranson mar 76

77 Gbbs: Spdng Convrgnc Rduc dpndnc bwn sampls auocorrlaon Skp sampls Randomz Varabl Samplng Ordr Employ blockng groupng Mulpl chans Rduc varanc covr n h n scon 77

78 Blockng Gbbs Samplr Sampl svral varabls oghr as a block Eampl: Gvn hr varabls YZ wh domans of sz group Y and Z oghr o form a varabl W={YZ} wh doman sz 4. Thn gvn sampl y z compu n sampl: y z w y z w Y Z + Can mprov convrgnc graly whn wo varabls ar srongly corrlad! - Doman of h block varabl grows ponnally wh h #varabls n a block! 78

79 Gbbs: Mulpl Chans Gnra M chans of sz K Each chan producs ndpndn sma m : 79 M m M ˆ K m K \ Tra m as ndpndn random varabls. Esma as avrag of m :

80 Gbbs Samplng Summary Markov Chan Mon Carlo mhod Glfand and Smh 990 Smh and Robrs 993 Trny 994 Sampls ar dpndn form Markov Chan Sampl from whch convrgs o Guarand o convrg whn all > 0 Mhods o mprov convrgnc: Blockng Rao-Blackwllsd 80

81 Ovrvw. robablsc Rasonng/Graphcal modls. Imporanc Samplng 3. Markov Chan Mon Carlo: Gbbs Samplng 4. Samplng n prsnc of Drmnsm 5. Rao-Blackwllsaon 6. AND/OR mporanc samplng

82 Samplng: rformanc Gbbs samplng Rduc dpndnc bwn sampls Imporanc samplng Rduc varanc Achv boh by samplng a subs of varabls and ngrang ou h rs rduc dmnsonaly aka Rao-Blackwllsaon Eplo graph srucur o manag h ra cos 8

83 Smallr Subs Sa-Spac Smallr sa-spac s asr o covr { 3 4} { } D 64 D 6 83

84 Smoohr Dsrbuon

85 Spdng Up Convrgnc Man Squard Error of h smaor: MSE Q BIAS Var Q In cas of unbasd smaor BIAS=0 MSE Q ˆ [ ˆ] VarQ [ ˆ] EQ EQ[ ] Rduc varanc spd up convrgnc! 85

86 Rao-Blackwllsaon 86 } { ~ ]} [ { } { } { ˆ ]} [ { } { ]} {var[ ]} [ { } { ]} [ ] [ { ~ } { ˆ g Var T l h E Var T h Var g Var l g E Var g Var l g E l g E Var g Var l h E l h E T g h h T g L R T T Lu Ch..3

87 Rao-Blackwllsaon Carry ou analycal compuaon as much as possbl - Lu =RL Imporanc Samplng: Var Q R L { } Var Q R L Q R { } Q R Lu Ch..5.5 Gbbs Samplng: auocovarancs ar lowr lss corrlaon bwn sampls f and j ar srongly corrlad =0 j =0 only nclud on fo hm no a samplng s 87

88 Blockng Gbbs Samplr vs. Collapsd Y Z Fasr Convrgnc Sandard Gbbs: y z y z z y Blockng: y z y z Collapsd: y y 3 88

89 Collapsd Gbbs Samplng Gnrang Sampls Gnra sampl c + from c : 89 \ c c c c C c c c c c C c c c c c C c c c c c C K K K K K K sampldfrom In shor for = o K:

90 Collapsd Gbbs Samplr Inpu: C E= Oupu: T sampls {c } F vdnc E= nalz c 0 a random. For = o T compu sampls. For = o N loop hrough varabls 3. c + C c \c 4. End For 5. End For

91 Calculaon Tm Compung c c \c s mor pnsv rqurs nfrnc Tradng #sampls for smallr varanc: gnra mor sampls wh hghr covaranc gnra fwr sampls wh lowr covaranc Mus conrol h m spn compung samplng probabls n ordr o b mffcv! 9

92 Eplong Graph roprs Rcall compuaon m s ponnal n h adjusd nducd wdh of a graph w-cus s a subs of varabl s.. whn hy ar obsrvd nducd wdh of h graph s w whn sampld varabls form a w-cus nfrnc s pw.g. usng Buck Tr Elmnaon cycl-cus s a spcal cas of w-cus Samplng w-cus w-cus samplng! 9

93 Wha If C=Cycl-Cus? c { 5 } E { 9} 5 9 can compu usng Buck Elmnaon compuaon comply s ON 93

94 Compung Transon robabls 94 : 0 : BE BE Compu jon probabls: Normalz:

95 Cus Samplng-Answrng Qurs Qury: c C c =? sam as Gbbs: ˆ c T T c c \ c compud whl gnrang sampl usng buck r lmnaon Qury: \C =? T c T compu afr gnrang sampl usng buck r lmnaon 95

96 Cus Samplng vs. Cus Condonng 96 c c T c coun c c T C D c C D c T Cus Condonng Cus Samplng c c C D c

97 Cus Samplng Eampl Esmang for samplng nod : Sampl Sampl Sampl

98 Cus Samplng Eampl 98 } { } { } { c c c Esmang 3 for non-sampld nod 3 :

99 CCS54 Ts Rsuls CCS54 n=54 C=5 E=3 CCS54 n=54 C=5 E= Cus Gbbs Cus Gbbs # sampls Tmsc MSE vs. #sampls lf and m rgh Ergodc =54 D = C=5 E=3 Eac Tm = 30 sc usng Cus Condonng 99

100 CCS79 Ts Rsuls CCS79 n=79 C=8 E=35 CCS79 n=79 C=8 E= Cus Gbbs Cus Gbbs # sampls Tmsc MSE vs. #sampls lf and m rgh Non-Ergodc drmnsc CT nry = 79 C = 8 <= D <=4 E = 35 Eac Tm = sc usng Cus Condonng 00

101 CCS360b Ts Rsuls CCS360b n=360 C= E=36 CCS360b n=360 C= E=36 Cus Gbbs Cus Gbbs # sampls Tmsc MSE vs. #sampls lf and m rgh Ergodc = 360 D = C = E = 36 Eac Tm > 60 mn usng Cus Condonng Eac Valus oband va Buck Elmnaon 0

102 Random Nworks RANDOM n=00 C=3 E=5-0 Cus Gbbs # sampls RANDOM n=00 C=3 E=5-0 Cus Gbbs Tmsc MSE vs. #sampls lf and m rgh = 00 D =C = 3 E = 5-0 Eac Tm = 30 sc usng Cus Condonng 0

103 Codng Nworks Cus Transforms Non-Ergodc Chan o Ergodc 3 4 Codng Nworks n=00 C=-4 u u u 3 u 4 0. IB Gbbs Cus p p p 3 p y y y 3 y Tmsc MSE vs. m rgh Non-Ergodc = 00 D = C = 3-6 E = 50 Sampl Ergodc Subspac U={U U U k } Eac Tm = 50 sc usng Cus Condonng 03

104 Non-Ergodc Halfndr HalFndr n=56 C=5 E= HalFndr n=56 C=5 E= 0. Cus Gbbs Cus Gbbs # sampls Tmsc MSE vs. #sampls lf and m rgh Non-Ergodc = 56 C = 5 <=D <= E = 0 Eac Tm = sc usng Loop-Cus Condonng 04

105 CCS360b - MSE cpcs360b N=360 E=[0-34] w*=0 MSE Gbbs IB C=6fw=3 C=48fw= Tm sc MSE vs. Tm Ergodc = 360 C = 6 D = Eac Tm = 50 mn usng BTE 05

106 Cus Imporanc Samplng Apply Imporanc Samplng ovr cus C T T w T Q c c T ˆ T w c c T c T w c T whr c s compud usng Buck Elmnaon Goga & Dchr 005 and Bdyuk & Dchr 006

107 Lklhood Cus Wghng LCS Z=Topologcal Ordr{CE} Gnrang sampl +: For If Z Z z Els z End If End For Z E do : z z Z z... z KL[C QC+ KL* Q] compud whl gnrang sampl usng buck r lmnaon can b mmozd for som numbr of nsancs K basd on mmory avalabl 07

108 ahfndr 08

109 ahfndr 09

110 Lnk 0

111 Summary Imporanc Samplng..d. sampls Unbasd smaor Gnras sampls fas Sampls from Q Rjc sampls wh zro-wgh Improvs on cus Gbbs Samplng Dpndn sampls Basd smaor Gnras sampls slowr Sampls from ` Dos no convrg n prsnc of consrans Improvs on cus

112 MSE CCS360b.E-0.E-03 cpcs360b N=360 LC=6 w*= E=5 LW AIS-BN Gbbs LCS IB.E-04.E Tm sc LW lklhood wghng LCS lklhood wghng on a cus

113 MSE CCS4b.0E-0.0E-03 cpcs4b N=4 LC=47 w*= E=8 LW AIS-BN Gbbs LCS IB.0E-04.0E Tm sc LW lklhood wghng LCS lklhood wghng on a cus 3

114 MSE Codng Nworks.0E-0.0E-0 codng N=00 =3 LC=6 w*= LW AIS-BN Gbbs LCS IB.0E-03.0E-04.0E Tm sc LW lklhood wghng LCS lklhood wghng on a cus 4

innovations shocks white noise

innovations shocks white noise Innovaons Tm-srs modls ar consrucd as lnar funcons of fundamnal forcasng rrors, also calld nnovaons or shocks Ths basc buldng blocks sasf var σ Srall uncorrlad Ths rrors ar calld wh nos In gnral, f ou

More information

Boosting and Ensemble Methods

Boosting and Ensemble Methods Boosng and Ensmbl Mhods PAC Larnng modl Som dsrbuon D ovr doman X Eampls: c* s h arg funcon Goal: Wh hgh probably -d fnd h n H such ha rrorh,c* < d and ar arbrarly small. Inro o ML 2 Wak Larnng

More information

Consider a system of 2 simultaneous first order linear equations

Consider a system of 2 simultaneous first order linear equations Soluon of sysms of frs ordr lnar quaons onsdr a sysm of smulanous frs ordr lnar quaons a b c d I has h alrna mar-vcor rprsnaon a b c d Or, n shorhand A, f A s alrady known from con W know ha h abov sysm

More information

t=0 t>0: + vr - i dvc Continuation

t=0 t>0: + vr - i dvc Continuation hapr Ga Dlay and rcus onnuaon s rcu Equaon >: S S Ths dffrnal quaon, oghr wh h nal condon, fully spcfs bhaor of crcu afr swch closs Our n challng: larn how o sol such quaons TUE/EE 57 nwrk analys 4/5 NdM

More information

The Variance-Covariance Matrix

The Variance-Covariance Matrix Th Varanc-Covaranc Marx Our bggs a so-ar has bn ng a lnar uncon o a s o daa by mnmzng h las squars drncs rom h o h daa wh mnsarch. Whn analyzng non-lnar daa you hav o us a program l Malab as many yps o

More information

Sampling Techniques for Probabilistic and Deterministic Graphical models

Sampling Techniques for Probabilistic and Deterministic Graphical models Samplng ehnques for robabls and Deermns Graphal models ICS 76 Fall 04 Bozhena Bdyuk Rna Deher Readng Darwhe haper 5 relaed papers Overvew. robabls Reasonng/Graphal models. Imporane Samplng 3. Markov Chan

More information

Advanced Queueing Theory. M/G/1 Queueing Systems

Advanced Queueing Theory. M/G/1 Queueing Systems Advand Quung Thory Ths slds ar rad by Dr. Yh Huang of Gorg Mason Unvrsy. Sudns rgsrd n Dr. Huang's ourss a GMU an ma a sngl mahn-radabl opy and prn a sngl opy of ah sld for hr own rfrn, so long as ah sld

More information

Lecture 4 : Backpropagation Algorithm. Prof. Seul Jung ( Intelligent Systems and Emotional Engineering Laboratory) Chungnam National University

Lecture 4 : Backpropagation Algorithm. Prof. Seul Jung ( Intelligent Systems and Emotional Engineering Laboratory) Chungnam National University Lcur 4 : Bacpropagaon Algorhm Pro. Sul Jung Inllgn Sm and moonal ngnrng Laboraor Chungnam Naonal Unvr Inroducon o Bacpropagaon algorhm 969 Mn and Papr aac. 980 Parr and Wrbo dcovrd bac propagaon algorhm.

More information

Summary: Solving a Homogeneous System of Two Linear First Order Equations in Two Unknowns

Summary: Solving a Homogeneous System of Two Linear First Order Equations in Two Unknowns Summary: Solvng a Homognous Sysm of Two Lnar Frs Ordr Equaons n Two Unknowns Gvn: A Frs fnd h wo gnvalus, r, and hr rspcv corrspondng gnvcors, k, of h coffcn mar A Dpndng on h gnvalus and gnvcors, h gnral

More information

Supplementary Figure 1. Experiment and simulation with finite qudit. anharmonicity. (a), Experimental data taken after a 60 ns three-tone pulse.

Supplementary Figure 1. Experiment and simulation with finite qudit. anharmonicity. (a), Experimental data taken after a 60 ns three-tone pulse. Supplmnar Fgur. Eprmn and smulaon wh fn qud anharmonc. a, Eprmnal daa akn afr a 6 ns hr-on puls. b, Smulaon usng h amlonan. Supplmnar Fgur. Phagoran dnamcs n h m doman. a, Eprmnal daa. Th hr-on puls s

More information

Homework: Introduction to Motion

Homework: Introduction to Motion Homwork: Inroducon o Moon Dsanc vs. Tm Graphs Nam Prod Drcons: Answr h foowng qusons n h spacs provdd. 1. Wha do you do o cra a horzona n on a dsancm graph? 2. How do you wak o cra a sragh n ha sops up?

More information

Wave Superposition Principle

Wave Superposition Principle Physcs 36: Was Lcur 5 /7/8 Wa Suroson Prncl I s qu a common suaon for wo or mor was o arr a h sam on n sac or o xs oghr along h sam drcon. W wll consdr oday sral moran cass of h combnd ffcs of wo or mor

More information

ELEN E4830 Digital Image Processing

ELEN E4830 Digital Image Processing ELEN E48 Dgal Imag Procssng Mrm Eamnaon Sprng Soluon Problm Quanzaon and Human Encodng r k u P u P u r r 6 6 6 6 5 6 4 8 8 4 P r 6 6 P r 4 8 8 6 8 4 r 8 4 8 4 7 8 r 6 6 6 6 P r 8 4 8 P r 6 6 8 5 P r /

More information

Problem 1: Consider the following stationary data generation process for a random variable y t. e t ~ N(0,1) i.i.d.

Problem 1: Consider the following stationary data generation process for a random variable y t. e t ~ N(0,1) i.i.d. A/CN C m Sr Anal Profor Òcar Jordà Wnr conomc.c. Dav POBLM S SOLIONS Par I Analcal Quon Problm : Condr h followng aonar daa gnraon proc for a random varabl - N..d. wh < and N -. a Oban h populaon man varanc

More information

CSE 245: Computer Aided Circuit Simulation and Verification

CSE 245: Computer Aided Circuit Simulation and Verification CSE 45: Compur Aidd Circui Simulaion and Vrificaion Fall 4, Sp 8 Lcur : Dynamic Linar Sysm Oulin Tim Domain Analysis Sa Equaions RLC Nwork Analysis by Taylor Expansion Impuls Rspons in im domain Frquncy

More information

9. Simple Rules for Monetary Policy

9. Simple Rules for Monetary Policy 9. Smpl Ruls for Monar Polc John B. Talor, Ma 0, 03 Woodford, AR 00 ovrvw papr Purpos s o consdr o wha xn hs prscrpon rsmbls h sor of polc ha conomc hor would rcommnd Bu frs, l s rvw how hs sor of polc

More information

Theoretical Seismology

Theoretical Seismology Thorcal Ssmology Lcur 9 Sgnal Procssng Fourr analyss Fourr sudd a h Écol Normal n Pars, augh by Lagrang, who Fourr dscrbd as h frs among Europan mn of scnc, Laplac, who Fourr rad lss hghly, and by Mong.

More information

State Observer Design

State Observer Design Sa Obsrvr Dsgn A. Khak Sdgh Conrol Sysms Group Faculy of Elcrcal and Compur Engnrng K. N. Toos Unvrsy of Tchnology Fbruary 2009 1 Problm Formulaon A ky assumpon n gnvalu assgnmn and sablzng sysms usng

More information

EE243 Advanced Electromagnetic Theory Lec # 10: Poynting s Theorem, Time- Harmonic EM Fields

EE243 Advanced Electromagnetic Theory Lec # 10: Poynting s Theorem, Time- Harmonic EM Fields Appl M Fall 6 Nuruhr Lcur # r 9/6/6 4 Avanc lcromagnc Thory Lc # : Poynng s Thorm Tm- armonc M Fls Poynng s Thorm Consrvaon o nrgy an momnum Poynng s Thorm or Lnar sprsv Ma Poynng s Thorm or Tm-armonc

More information

FAULT TOLERANT SYSTEMS

FAULT TOLERANT SYSTEMS FAULT TOLERANT SYSTEMS hp://www.cs.umass.du/c/orn/faultolransysms ar 4 Analyss Mhods Chapr HW Faul Tolranc ar.4.1 Duplx Sysms Boh procssors xcu h sam as If oupus ar n agrmn - rsul s assumd o b corrc If

More information

On the Derivatives of Bessel and Modified Bessel Functions with Respect to the Order and the Argument

On the Derivatives of Bessel and Modified Bessel Functions with Respect to the Order and the Argument Inrnaional Rsarch Journal of Applid Basic Scincs 03 Aailabl onlin a wwwirjabscom ISSN 5-838X / Vol 4 (): 47-433 Scinc Eplorr Publicaions On h Driais of Bssl Modifid Bssl Funcions wih Rspc o h Ordr h Argumn

More information

Chapter 13 Laplace Transform Analysis

Chapter 13 Laplace Transform Analysis Chapr aplac Tranorm naly Chapr : Ouln aplac ranorm aplac Tranorm -doman phaor analy: x X σ m co ω φ x X X m φ x aplac ranorm: [ o ] d o d < aplac Tranorm Thr condon Unlaral on-dd aplac ranorm: aplac ranorm

More information

SIMEON BALL AND AART BLOKHUIS

SIMEON BALL AND AART BLOKHUIS A BOUND FOR THE MAXIMUM WEIGHT OF A LINEAR CODE SIMEON BALL AND AART BLOKHUIS Absrac. I s shown ha h paramrs of a lnar cod ovr F q of lngh n, dmnson k, mnmum wgh d and maxmum wgh m sasfy a cran congrunc

More information

AR(1) Process. The first-order autoregressive process, AR(1) is. where e t is WN(0, σ 2 )

AR(1) Process. The first-order autoregressive process, AR(1) is. where e t is WN(0, σ 2 ) AR() Procss Th firs-ordr auorgrssiv procss, AR() is whr is WN(0, σ ) Condiional Man and Varianc of AR() Condiional man: Condiional varianc: ) ( ) ( Ω Ω E E ) var( ) ) ( var( ) var( σ Ω Ω Ω Ω E Auocovarianc

More information

Dr. Junchao Xia Center of Biophysics and Computational Biology. Fall /21/2016 1/23

Dr. Junchao Xia Center of Biophysics and Computational Biology. Fall /21/2016 1/23 BIO53 Bosascs Lcur 04: Cral Lm Thorm ad Thr Dsrbuos Drvd from h Normal Dsrbuo Dr. Juchao a Cr of Bophyscs ad Compuaoal Bology Fall 06 906 3 Iroduco I hs lcur w wll alk abou ma cocps as lsd blow, pcd valu

More information

Frequency Response. Response of an LTI System to Eigenfunction

Frequency Response. Response of an LTI System to Eigenfunction Frquncy Rsons Las m w Rvsd formal dfnons of lnary and m-nvaranc Found an gnfuncon for lnar m-nvaran sysms Found h frquncy rsons of a lnar sysm o gnfuncon nu Found h frquncy rsons for cascad, fdbac, dffrnc

More information

Poisson process Markov process

Poisson process Markov process E2200 Quuing hory and lraffic 2nd lcur oion proc Markov proc Vikoria Fodor KTH Laboraory for Communicaion nwork, School of Elcrical Enginring 1 Cour oulin Sochaic proc bhind quuing hory L2-L3 oion proc

More information

Review - Probabilistic Classification

Review - Probabilistic Classification Mmoral Unvrsty of wfoundland Pattrn Rcognton Lctur 8 May 5, 6 http://www.ngr.mun.ca/~charlsr Offc Hours: Tusdays Thursdays 8:3-9:3 PM E- (untl furthr notc) Gvn lablld sampls { ɛc,,,..., } {. Estmat Rvw

More information

8. Queueing systems. Contents. Simple teletraffic model. Pure queueing system

8. Queueing systems. Contents. Simple teletraffic model. Pure queueing system 8. Quug sysms Cos 8. Quug sysms Rfrshr: Sml lraffc modl Quug dscl M/M/ srvr wag lacs Alcao o ack lvl modllg of daa raffc M/M/ srvrs wag lacs lc8. S-38.45 Iroduco o Tlraffc Thory Srg 5 8. Quug sysms 8.

More information

UNIT #5 EXPONENTIAL AND LOGARITHMIC FUNCTIONS

UNIT #5 EXPONENTIAL AND LOGARITHMIC FUNCTIONS Answr Ky Nam: Da: UNIT # EXPONENTIAL AND LOGARITHMIC FUNCTIONS Par I Qusions. Th prssion is quivaln o () () 6 6 6. Th ponnial funcion y 6 could rwrin as y () y y 6 () y y (). Th prssion a is quivaln o

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

CPSC 211 Data Structures & Implementations (c) Texas A&M University [ 259] B-Trees

CPSC 211 Data Structures & Implementations (c) Texas A&M University [ 259] B-Trees CPSC 211 Daa Srucurs & Implmnaions (c) Txas A&M Univrsiy [ 259] B-Trs Th AVL r and rd-black r allowd som variaion in h lnghs of h diffrn roo-o-laf pahs. An alrnaiv ida is o mak sur ha all roo-o-laf pahs

More information

Final Exam : Solutions

Final Exam : Solutions Comp : Algorihm and Daa Srucur Final Exam : Soluion. Rcuriv Algorihm. (a) To bgin ind h mdian o {x, x,... x n }. Sinc vry numbr xcp on in h inrval [0, n] appar xacly onc in h li, w hav ha h mdian mu b

More information

Chapter 9 Transient Response

Chapter 9 Transient Response har 9 Transn sons har 9: Ouln N F n F Frs-Ordr Transns Frs-Ordr rcus Frs ordr crcus: rcus conan onl on nducor or on caacor gornd b frs-ordr dffrnal quaons. Zro-nu rsons: h crcu has no ald sourc afr a cran

More information

Microscopic Flow Characteristics Time Headway - Distribution

Microscopic Flow Characteristics Time Headway - Distribution CE57: Traffic Flow Thory Spring 20 Wk 2 Modling Hadway Disribuion Microscopic Flow Characrisics Tim Hadway - Disribuion Tim Hadway Dfiniion Tim Hadway vrsus Gap Ahmd Abdl-Rahim Civil Enginring Dparmn,

More information

Lecture 1: Numerical Integration The Trapezoidal and Simpson s Rule

Lecture 1: Numerical Integration The Trapezoidal and Simpson s Rule Lcur : Numrical ngraion Th Trapzoidal and Simpson s Rul A problm Th probabiliy of a normally disribud (man µ and sandard dviaion σ ) vn occurring bwn h valus a and b is B A P( a x b) d () π whr a µ b -

More information

EXERCISE - 01 CHECK YOUR GRASP

EXERCISE - 01 CHECK YOUR GRASP DIFFERENTIAL EQUATION EXERCISE - CHECK YOUR GRASP 7. m hn D() m m, D () m m. hn givn D () m m D D D + m m m m m m + m m m m + ( m ) (m ) (m ) (m + ) m,, Hnc numbr of valus of mn will b. n ( ) + c sinc

More information

Almost unbiased exponential estimator for the finite population mean

Almost unbiased exponential estimator for the finite population mean Almos ubasd poal smaor for f populao ma Rajs Sg, Pakaj aua, ad rmala Saa, Scool of Sascs, DAVV, Idor (M.P., Ida (rsgsa@aoo.com Flor Smaradac ar of Dparm of Mamacs, Uvrs of Mco, Gallup, USA (smarad@um.du

More information

Lucas Test is based on Euler s theorem which states that if n is any integer and a is coprime to n, then a φ(n) 1modn.

Lucas Test is based on Euler s theorem which states that if n is any integer and a is coprime to n, then a φ(n) 1modn. Modul 10 Addtonal Topcs 10.1 Lctur 1 Prambl: Dtrmnng whthr a gvn ntgr s prm or compost s known as prmalty tstng. Thr ar prmalty tsts whch mrly tll us whthr a gvn ntgr s prm or not, wthout gvng us th factors

More information

Institute of Actuaries of India

Institute of Actuaries of India Insiu of Acuaris of India ubjc CT3 Probabiliy and Mahmaical aisics Novmbr Examinaions INDICATIVE OLUTION Pag of IAI CT3 Novmbr ol. a sampl man = 35 sampl sandard dviaion = 36.6 b for = uppr bound = 35+*36.6

More information

Chap 2: Reliability and Availability Models

Chap 2: Reliability and Availability Models Chap : lably ad valably Modls lably = prob{s s fully fucog [,]} Suppos from [,] m prod, w masur ou of N compos, of whch N : # of compos oprag corrcly a m N f : # of compos whch hav fald a m rlably of h

More information

DEPARTMENT OF ELECTRICAL &ELECTRONICS ENGINEERING SIGNALS AND SYSTEMS. Assoc. Prof. Dr. Burak Kelleci. Spring 2018

DEPARTMENT OF ELECTRICAL &ELECTRONICS ENGINEERING SIGNALS AND SYSTEMS. Assoc. Prof. Dr. Burak Kelleci. Spring 2018 DEPARTMENT OF ELECTRICAL &ELECTRONICS ENGINEERING SIGNALS AND SYSTEMS Aoc. Prof. Dr. Burak Kllci Spring 08 OUTLINE Th Laplac Tranform Rgion of convrgnc for Laplac ranform Invr Laplac ranform Gomric valuaion

More information

Basic Polyhedral theory

Basic Polyhedral theory Basic Polyhdral thory Th st P = { A b} is calld a polyhdron. Lmma 1. Eithr th systm A = b, b 0, 0 has a solution or thr is a vctorπ such that π A 0, πb < 0 Thr cass, if solution in top row dos not ist

More information

Applying Software Reliability Techniques to Low Retail Demand Estimation

Applying Software Reliability Techniques to Low Retail Demand Estimation Applyng Sofwar Rlably Tchnqus o Low Ral Dmand Esmaon Ma Lndsy Unvrsy of Norh Txas ITDS Dp P.O. Box 30549 Dnon, TX 7603-549 940 565 3174 lndsym@un.du Robr Pavur Unvrsy of Norh Txas ITDS Dp P.O. Box 30549

More information

The Hyperelastic material is examined in this section.

The Hyperelastic material is examined in this section. 4. Hyprlastcty h Hyprlastc matral s xad n ths scton. 4..1 Consttutv Equatons h rat of chang of ntrnal nrgy W pr unt rfrnc volum s gvn by th strss powr, whch can b xprssd n a numbr of dffrnt ways (s 3.7.6):

More information

Conventional Hot-Wire Anemometer

Conventional Hot-Wire Anemometer Convnonal Ho-Wr Anmomr cro Ho Wr Avanag much mallr prob z mm o µm br paal roluon array o h nor hghr rquncy rpon lowr co prormanc/co abrcaon roc I µm lghly op p layr 8µm havly boron op ch op layr abrcaon

More information

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model

Outline. Probabilistic Model Learning. Probabilistic Model Learning. Probabilistic Model for Time-series Data: Hidden Markov Model Probablsc Model for Tme-seres Daa: Hdden Markov Model Hrosh Mamsuka Bonformacs Cener Kyoo Unversy Oulne Three Problems for probablsc models n machne learnng. Compung lkelhood 2. Learnng 3. Parsng (predcon

More information

Elementary Differential Equations and Boundary Value Problems

Elementary Differential Equations and Boundary Value Problems Elmnar Diffrnial Equaions and Boundar Valu Problms Boc. & DiPrima 9 h Ediion Chapr : Firs Ordr Diffrnial Equaions 00600 คณ ตศาสตร ว ศวกรรม สาขาว ชาว ศวกรรมคอมพ วเตอร ป การศ กษา /55 ผศ.ดร.อร ญญา ผศ.ดร.สมศ

More information

CIVL 8/ D Boundary Value Problems - Triangular Elements (T6) 1/8

CIVL 8/ D Boundary Value Problems - Triangular Elements (T6) 1/8 CIVL 8/7 -D Boundar Valu Problm - rangular Elmn () /8 SI-ODE RIAGULAR ELEMES () A quadracall nrpolad rangular lmn dfnd b nod, hr a h vrc and hr a h mddl a ach d. h mddl nod, dpndng on locaon, ma dfn a

More information

Search sequence databases 3 10/25/2016

Search sequence databases 3 10/25/2016 Sarch squnc databass 3 10/25/2016 Etrm valu distribution Ø Suppos X is a random variabl with probability dnsity function p(, w sampl a larg numbr S of indpndnt valus of X from this distribution for an

More information

Spring 2006 Process Dynamics, Operations, and Control Lesson 2: Mathematics Review

Spring 2006 Process Dynamics, Operations, and Control Lesson 2: Mathematics Review Spring 6 Procss Dynamics, Opraions, and Conrol.45 Lsson : Mahmaics Rviw. conx and dircion Imagin a sysm ha varis in im; w migh plo is oupu vs. im. A plo migh imply an quaion, and h quaion is usually an

More information

4.1 The Uniform Distribution Def n: A c.r.v. X has a continuous uniform distribution on [a, b] when its pdf is = 1 a x b

4.1 The Uniform Distribution Def n: A c.r.v. X has a continuous uniform distribution on [a, b] when its pdf is = 1 a x b 4. Th Uniform Disribuion Df n: A c.r.v. has a coninuous uniform disribuion on [a, b] whn is pdf is f x a x b b a Also, b + a b a µ E and V Ex4. Suppos, h lvl of unblivabiliy a any poin in a Transformrs

More information

Transient Analysis of Two-dimensional State M/G/1 Queueing Model with Multiple Vacations and Bernoulli Schedule

Transient Analysis of Two-dimensional State M/G/1 Queueing Model with Multiple Vacations and Bernoulli Schedule Inrnaonal Journal of Compur Applcaons (975 8887) Volum 4 No.3, Fbruary 22 Transn Analyss of Two-dmnsonal Sa M/G/ Quung Modl wh Mulpl Vacaons and Brnoull Schdul Indra Assoca rofssor Dparmn of Sascs and

More information

1. Inverse Matrix 4[(3 7) (02)] 1[(0 7) (3 2)] Recall that the inverse of A is equal to:

1. Inverse Matrix 4[(3 7) (02)] 1[(0 7) (3 2)] Recall that the inverse of A is equal to: Rfrncs Brnank, B. and I. Mihov (1998). Masuring monary policy, Quarrly Journal of Economics CXIII, 315-34. Blanchard, O. R. Proi (00). An mpirical characrizaion of h dynamic ffcs of changs in govrnmn spnding

More information

S.Y. B.Sc. (IT) : Sem. III. Applied Mathematics. Q.1 Attempt the following (any THREE) [15]

S.Y. B.Sc. (IT) : Sem. III. Applied Mathematics. Q.1 Attempt the following (any THREE) [15] S.Y. B.Sc. (IT) : Sm. III Applid Mahmaics Tim : ½ Hrs.] Prlim Qusion Papr Soluion [Marks : 75 Q. Amp h following (an THREE) 3 6 Q.(a) Rduc h mari o normal form and find is rank whr A 3 3 5 3 3 3 6 Ans.:

More information

Control Systems (Lecture note #6)

Control Systems (Lecture note #6) 6.5 Corol Sysms (Lcur o #6 Las Tm: Lar algbra rw Lar algbrac quaos soluos Paramrzao of all soluos Smlary rasformao: compao form Egalus ad gcors dagoal form bg pcur: o brach of h cours Vcor spacs marcs

More information

A Note on Estimability in Linear Models

A Note on Estimability in Linear Models Intrnatonal Journal of Statstcs and Applcatons 2014, 4(4): 212-216 DOI: 10.5923/j.statstcs.20140404.06 A Not on Estmablty n Lnar Modls S. O. Adymo 1,*, F. N. Nwob 2 1 Dpartmnt of Mathmatcs and Statstcs,

More information

CHAPTER CHAPTER14. Expectations: The Basic Tools. Prepared by: Fernando Quijano and Yvonn Quijano

CHAPTER CHAPTER14. Expectations: The Basic Tools. Prepared by: Fernando Quijano and Yvonn Quijano Expcaions: Th Basic Prpard by: Frnando Quijano and Yvonn Quijano CHAPTER CHAPTER14 2006 Prnic Hall Businss Publishing Macroconomics, 4/ Olivir Blanchard 14-1 Today s Lcur Chapr 14:Expcaions: Th Basic Th

More information

Least squares and motion. Nuno Vasconcelos ECE Department, UCSD

Least squares and motion. Nuno Vasconcelos ECE Department, UCSD Las squars ad moo uo Vascoclos ECE Dparm UCSD Pla for oda oda w wll dscuss moo smao hs s rsg wo was moo s vr usful as a cu for rcogo sgmao comprsso c. s a gra ampl of las squars problm w wll also wrap

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

Boyce/DiPrima 9 th ed, Ch 2.1: Linear Equations; Method of Integrating Factors

Boyce/DiPrima 9 th ed, Ch 2.1: Linear Equations; Method of Integrating Factors Boc/DiPrima 9 h d, Ch.: Linar Equaions; Mhod of Ingraing Facors Elmnar Diffrnial Equaions and Boundar Valu Problms, 9 h diion, b William E. Boc and Richard C. DiPrima, 009 b John Wil & Sons, Inc. A linar

More information

Engineering Circuit Analysis 8th Edition Chapter Nine Exercise Solutions

Engineering Circuit Analysis 8th Edition Chapter Nine Exercise Solutions Engnrng rcu naly 8h Eon hapr Nn Exrc Soluon. = KΩ, = µf, an uch ha h crcu rpon oramp. a For Sourc-fr paralll crcu: For oramp or b H 9V, V / hoo = H.7.8 ra / 5..7..9 9V 9..9..9 5.75,.5 5.75.5..9 . = nh,

More information

Gaussian Random Process and Its Application for Detecting the Ionospheric Disturbances Using GPS

Gaussian Random Process and Its Application for Detecting the Ionospheric Disturbances Using GPS Journal of Global Posonng Sysms (005) Vol. 4, No. 1-: 76-81 Gaussan Random Procss and Is Applcaon for Dcng h Ionosphrc Dsurbancs Usng GPS H.. Zhang 1,, J. Wang 3, W. Y. Zhu 1, C. Huang 1 (1) Shangha Asronomcal

More information

CHAPTER 2: Supervised Learning

CHAPTER 2: Supervised Learning HATER 2: Supervsed Learnng Learnng a lass from Eamples lass of a famly car redcon: Is car a famly car? Knowledge eracon: Wha do people epec from a famly car? Oupu: osve (+) and negave ( ) eamples Inpu

More information

Bethe-Salpeter Equation Green s Function and the Bethe-Salpeter Equation for Effective Interaction in the Ladder Approximation

Bethe-Salpeter Equation Green s Function and the Bethe-Salpeter Equation for Effective Interaction in the Ladder Approximation Bh-Salp Equaon n s Funcon and h Bh-Salp Equaon fo Effcv Inacon n h Ladd Appoxmaon Csa A. Z. Vasconcllos Insuo d Físca-UFRS - upo: Físca d Hadons Sngl-Pacl Popagao. Dagam xpanson of popagao. W consd as

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

Text: WMM, Chapter 5. Sections , ,

Text: WMM, Chapter 5. Sections , , Lcturs 6 - Continuous Probabilit Distributions Tt: WMM, Chaptr 5. Sctions 6.-6.4, 6.6-6.8, 7.-7. In th prvious sction, w introduc som of th common probabilit distribution functions (PDFs) for discrt sampl

More information

A L A BA M A L A W R E V IE W

A L A BA M A L A W R E V IE W A L A BA M A L A W R E V IE W Volume 52 Fall 2000 Number 1 B E F O R E D I S A B I L I T Y C I V I L R I G HT S : C I V I L W A R P E N S I O N S A N D TH E P O L I T I C S O F D I S A B I L I T Y I N

More information

Copyright 2012 Pearson Education, Inc. Publishing as Prentice Hall.

Copyright 2012 Pearson Education, Inc. Publishing as Prentice Hall. Chapr Rviw 0 6. ( a a ln a. This will qual a if an onl if ln a, or a. + k an (ln + c. Thrfor, a an valu of, whr h wo curvs inrsc, h wo angn lins will b prpnicular. 6. (a Sinc h lin passs hrough h origin

More information

Analysis of decentralized potential field based multi-agent navigation via primal-dual Lyapunov theory

Analysis of decentralized potential field based multi-agent navigation via primal-dual Lyapunov theory Analyss of dcnralzd ponal fld basd mul-agn navgaon va prmal-dual Lyapunov hory Th MIT Faculy has mad hs arcl opnly avalabl. Plas shar how hs accss bnfs you. Your sory mars. Caon As Publshd Publshr Dmarogonas,

More information

Solution: APPM 1360 Final (150 pts) Spring (60 pts total) The following parts are not related, justify your answers:

Solution: APPM 1360 Final (150 pts) Spring (60 pts total) The following parts are not related, justify your answers: APPM 6 Final 5 pts) Spring 4. 6 pts total) Th following parts ar not rlatd, justify your answrs: a) Considr th curv rprsntd by th paramtric quations, t and y t + for t. i) 6 pts) Writ down th corrsponding

More information

Transfer function and the Laplace transformation

Transfer function and the Laplace transformation Lab No PH-35 Porland Sa Univriy A. La Roa Tranfr funcion and h Laplac ranformaion. INTRODUTION. THE LAPLAE TRANSFORMATION L 3. TRANSFER FUNTIONS 4. ELETRIAL SYSTEMS Analyi of h hr baic paiv lmn R, and

More information

Chapter 7 Stead St y- ate Errors

Chapter 7 Stead St y- ate Errors Char 7 Say-Sa rror Inroucon Conrol ym analy an gn cfcaon a. rann ron b. Sably c. Say-a rror fnon of ay-a rror : u c a whr u : nu, c: ouu Val only for abl ym chck ym ably fr! nu for ay-a a nu analy U o

More information

Continuous probability distributions

Continuous probability distributions Continuous probability distributions Many continuous probability distributions, including: Uniform Normal Gamma Eponntial Chi-Squard Lognormal Wibull EGR 5 Ch. 6 Uniform distribution Simplst charactrizd

More information

I) Title: Rational Expectations and Adaptive Learning. II) Contents: Introduction to Adaptive Learning

I) Title: Rational Expectations and Adaptive Learning. II) Contents: Introduction to Adaptive Learning I) Til: Raional Expcaions and Adapiv Larning II) Conns: Inroducion o Adapiv Larning III) Documnaion: - Basdvan, Olivir. (2003). Larning procss and raional xpcaions: an analysis using a small macroconomic

More information

10/7/14. Mixture Models. Comp 135 Introduction to Machine Learning and Data Mining. Maximum likelihood estimation. Mixture of Normals in 1D

10/7/14. Mixture Models. Comp 135 Introduction to Machine Learning and Data Mining. Maximum likelihood estimation. Mixture of Normals in 1D Comp 35 Introducton to Machn Larnng and Data Mnng Fall 204 rofssor: Ron Khardon Mxtur Modls Motvatd by soft k-mans w dvlopd a gnratv modl for clustrng. Assum thr ar k clustrs Clustrs ar not rqurd to hav

More information

Improved Exponential Estimator for Population Variance Using Two Auxiliary Variables

Improved Exponential Estimator for Population Variance Using Two Auxiliary Variables Improvd Epoal Emaor for Populao Varac Ug Two Aular Varabl Rajh gh Dparm of ac,baara Hdu Uvr(U.P., Ida (rgha@ahoo.com Pakaj Chauha ad rmala awa chool of ac, DAVV, Idor (M.P., Ida Flor maradach Dparm of

More information

Cramér-Rao Inequality: Let f(x; θ) be a probability density function with continuous parameter

Cramér-Rao Inequality: Let f(x; θ) be a probability density function with continuous parameter WHEN THE CRAMÉR-RAO INEQUALITY PROVIDES NO INFORMATION STEVEN J. MILLER Abstract. W invstigat a on-paramtr family of probability dnsitis (rlatd to th Parto distribution, which dscribs many natural phnomna)

More information

SECTION where P (cos θ, sin θ) and Q(cos θ, sin θ) are polynomials in cos θ and sin θ, provided Q is never equal to zero.

SECTION where P (cos θ, sin θ) and Q(cos θ, sin θ) are polynomials in cos θ and sin θ, provided Q is never equal to zero. SETION 6. 57 6. Evaluation of Dfinit Intgrals Exampl 6.6 W hav usd dfinit intgrals to valuat contour intgrals. It may com as a surpris to larn that contour intgrals and rsidus can b usd to valuat crtain

More information

Physics 160 Lecture 3. R. Johnson April 6, 2015

Physics 160 Lecture 3. R. Johnson April 6, 2015 Physics 6 Lcur 3 R. Johnson April 6, 5 RC Circui (Low-Pass Filr This is h sam RC circui w lookd a arlir h im doma, bu hr w ar rsd h frquncy rspons. So w pu a s wav sad of a sp funcion. whr R C RC Complx

More information

arxiv: v1 [math.ap] 16 Apr 2016

arxiv: v1 [math.ap] 16 Apr 2016 Th Cauchy problm for a combuson modl n porous mda J. C. da Moa M. M. Sanos. A. Sanos arxv:64.4798v [mah.ap] 6 Apr 6 Absrac W prov h xsnc of a global soluon o h Cauchy problm for a nonlnar racon-dffuson

More information

Economics 302 (Sec. 001) Intermediate Macroeconomic Theory and Policy (Spring 2011) 3/28/2012. UW Madison

Economics 302 (Sec. 001) Intermediate Macroeconomic Theory and Policy (Spring 2011) 3/28/2012. UW Madison Economics 302 (Sc. 001) Inrmdia Macroconomic Thory and Policy (Spring 2011) 3/28/2012 Insrucor: Prof. Mnzi Chinn Insrucor: Prof. Mnzi Chinn UW Madison 16 1 Consumpion Th Vry Forsighd dconsumr A vry forsighd

More information

Jones vector & matrices

Jones vector & matrices Jons vctor & matrcs PY3 Colást na hollscol Corcagh, Ér Unvrst Collg Cork, Irland Dpartmnt of Phscs Matr tratmnt of polarzaton Consdr a lght ra wth an nstantanous -vctor as shown k, t ˆ k, t ˆ k t, o o

More information

The Fourier Transform

The Fourier Transform /9/ Th ourr Transform Jan Baptst Josph ourr 768-83 Effcnt Data Rprsntaton Data can b rprsntd n many ways. Advantag usng an approprat rprsntaton. Eampls: osy ponts along a ln Color spac rd/grn/blu v.s.

More information

Random Process Part 1

Random Process Part 1 Random Procss Part A random procss t (, ζ is a signal or wavform in tim. t : tim ζ : outcom in th sampl spac Each tim w rapat th xprimnt, a nw wavform is gnratd. ( W will adopt t for short. Tim sampls

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

CHAPTER: 3 INVERSE EXPONENTIAL DISTRIBUTION: DIFFERENT METHOD OF ESTIMATIONS

CHAPTER: 3 INVERSE EXPONENTIAL DISTRIBUTION: DIFFERENT METHOD OF ESTIMATIONS CHAPTER: 3 INVERSE EXPONENTIAL DISTRIBUTION: DIFFERENT METHOD OF ESTIMATIONS 3. INTRODUCTION Th Ivrs Expoal dsrbuo was roducd by Kllr ad Kamah (98) ad has b sudd ad dscussd as a lfm modl. If a radom varabl

More information

2.1. Differential Equations and Solutions #3, 4, 17, 20, 24, 35

2.1. Differential Equations and Solutions #3, 4, 17, 20, 24, 35 MATH 5 PS # Summr 00.. Diffrnial Equaions and Soluions PS.# Show ha ()C #, 4, 7, 0, 4, 5 ( / ) is a gnral soluion of h diffrnial quaion. Us a compur or calculaor o skch h soluions for h givn valus of h

More information

On the Speed of Heat Wave. Mihály Makai

On the Speed of Heat Wave. Mihály Makai On h Spd of Ha Wa Mihály Maai maai@ra.bm.hu Conns Formulaion of h problm: infini spd? Local hrmal qulibrium (LTE hypohsis Balanc quaion Phnomnological balanc Spd of ha wa Applicaion in plasma ranspor 1.

More information

NAME: ANSWER KEY DATE: PERIOD. DIRECTIONS: MULTIPLE CHOICE. Choose the letter of the correct answer.

NAME: ANSWER KEY DATE: PERIOD. DIRECTIONS: MULTIPLE CHOICE. Choose the letter of the correct answer. R A T T L E R S S L U G S NAME: ANSWER KEY DATE: PERIOD PREAP PHYSICS REIEW TWO KINEMATICS / GRAPHING FORM A DIRECTIONS: MULTIPLE CHOICE. Chs h r f h rr answr. Us h fgur bw answr qusns 1 and 2. 0 10 20

More information

Grand Canonical Ensemble

Grand Canonical Ensemble Th nsmbl of systms mmrsd n a partcl-hat rsrvor at constant tmpratur T, prssur P, and chmcal potntal. Consdr an nsmbl of M dntcal systms (M =,, 3,...M).. Thy ar mutually sharng th total numbr of partcls

More information

Discussion 06 Solutions

Discussion 06 Solutions STAT Discussion Soluions Spring 8. Th wigh of fish in La Paradis follows a normal disribuion wih man of 8. lbs and sandard dviaion of. lbs. a) Wha proporion of fish ar bwn 9 lbs and lbs? æ 9-8. - 8. P

More information

Comparative Study of Finite Element and Haar Wavelet Correlation Method for the Numerical Solution of Parabolic Type Partial Differential Equations

Comparative Study of Finite Element and Haar Wavelet Correlation Method for the Numerical Solution of Parabolic Type Partial Differential Equations ISS 746-7659, England, UK Journal of Informaon and Compung Scnc Vol., o. 3, 6, pp.88-7 Comparav Sudy of Fn Elmn and Haar Wavl Corrlaon Mhod for h umrcal Soluon of Parabolc Typ Paral Dffrnal Equaons S.

More information

LECTURE 6 TRANSFORMATION OF RANDOM VARIABLES

LECTURE 6 TRANSFORMATION OF RANDOM VARIABLES LECTURE 6 TRANSFORMATION OF RANDOM VARIABLES TRANSFORMATION OF FUNCTION OF A RANDOM VARIABLE UNIVARIATE TRANSFORMATIONS TRANSFORMATION OF RANDOM VARIABLES If s a rv wth cdf F th Y=g s also a rv. If w wrt

More information

Linear Non-Gaussian Structural Equation Models

Linear Non-Gaussian Structural Equation Models IMPS 8, Durham, NH Linar Non-Gaussian Structural Equation Modls Shohi Shimizu, Patrik Hoyr and Aapo Hyvarinn Osaka Univrsity, Japan Univrsity of Hlsinki, Finland Abstract Linar Structural Equation Modling

More information

Mixture Ratio Estimators Using Multi-Auxiliary Variables and Attributes for Two-Phase Sampling

Mixture Ratio Estimators Using Multi-Auxiliary Variables and Attributes for Two-Phase Sampling Opn Journal of Sascs 04 4 776-788 Publshd Onln Ocobr 04 n Scs hp://scrporg/ournal/os hp://ddoorg/0436/os0449073 Mur ao Esmaors Usng Mul-Aular Varabls and Arbus for To-Phas Samplng Paul Mang Waru John Kung

More information

Boyce/DiPrima 9 th ed, Ch 7.8: Repeated Eigenvalues

Boyce/DiPrima 9 th ed, Ch 7.8: Repeated Eigenvalues Boy/DiPrima 9 h d Ch 7.8: Rpad Eignvalus Elmnary Diffrnial Equaions and Boundary Valu Problms 9 h diion by William E. Boy and Rihard C. DiPrima 9 by John Wily & Sons In. W onsidr again a homognous sysm

More information

Chapter 5 The Laplace Transform. x(t) input y(t) output Dynamic System

Chapter 5 The Laplace Transform. x(t) input y(t) output Dynamic System EE 422G No: Chapr 5 Inrucor: Chung Chapr 5 Th Laplac Tranform 5- Inroducion () Sym analyi inpu oupu Dynamic Sym Linar Dynamic ym: A procor which proc h inpu ignal o produc h oupu dy ( n) ( n dy ( n) +

More information

3(8 ) (8 x x ) 3x x (8 )

3(8 ) (8 x x ) 3x x (8 ) Scion - CHATER -. a d.. b. d.86 c d 8 d d.9997 f g 6. d. d. Thn, = ln. =. =.. d Thn, = ln.9 =.7 8 -. a d.6 6 6 6 6 8 8 8 b 9 d 6 6 6 8 c d.8 6 6 6 6 8 8 7 7 d 6 d.6 6 6 6 6 6 6 8 u u u u du.9 6 6 6 6 6

More information

Ma/CS 6a Class 15: Flows and Bipartite Graphs

Ma/CS 6a Class 15: Flows and Bipartite Graphs //206 Ma/CS 6a Cla : Flow and Bipari Graph By Adam Shffr Rmindr: Flow Nwork A flow nwork i a digraph G = V, E, oghr wih a ourc vrx V, a ink vrx V, and a capaciy funcion c: E N. Capaciy Sourc 7 a b c d

More information