Size: px
Start display at page:

Download ""

Transcription

1 A HIDDEN ARKOV ODEL APPROACH FOR LIHOLOGY IDENIFICAION FRO LOGS ara Padron, Sona Garca-Salce, Danel Barraez, Bernadee Dorzz, Sylve hra Insu Naonal des élécouncaons (IN, Evry, France; Unversdad Cenral de Venezuela (UCV, Caracas, Venezuela; LODYC, Unversé Perre e are Cure, Pars, France ara.padron@n-evry.fr, Sona.Salce@n-evry.fr, dbarraez@euler.cens.ucv.ve, Bernadee.Dorzz@nevry.fr, Sylve.hra@lodyc.jusseu.fr. INRODUCION We presen a new sascal ehod of denfyng lhologes relyng on wrelne log easureens ade on wo holes fro he french se of arcoule. Snce several years, sascal echnques have appeared as a powerful ool o classfy coplex and heerogeneous reservor lhology: ulvarae Sascs [Doveon (994], Dscrnan Analyss [Busch (987] and, ore recenly, Neural Neworks have been appled o hs proble. Concernng Neural echnques, ullayer Perceprons (LPs are used o classfy lhologes, eher relyng on well-logs drecly [Sauel (992], or soees afer usng Kohonen aps o deerne he lhologes of he reservor [Saggaf (2000]. Also, Self-Organzng aps have been used o reconsruc he lhologc faces of a drllng hole [Frayssne (2000, Anouar (997]. he goal of hs sudy s o denfy lhologes fro logs, relyng on nforaon abou rocks porosy and pereably. o hs end, we propose an orgnal approach based on Hdden arkov odels (Hs. Indeed, we consder a log sere of a drllng hole as a sequence of easures, and propose o odel n he sascal fraework gven by Hs. he reason s ha, n hs way, we can ake no accoun conexual dependences beween easures ade a dfferen levels of he drllng hole, whle perforng lhology denfcaon. In parcular, n coplex reservors, several lhologes are xured, and s exreely dffcul, even for a huan exper, o deerne whch s he lhology relyng only on he log easures aken a a gven level. In hs fraework, conexual nforaon ay be of porance o prove classfcaon a a gven level. Hs are ndeed wellknown sascal odels n oher applcave areas (lke speech recognon, on-lne handwrng recognon, ec... hey appear o be a powerful ool o explo conexual nforaon when perforng classfcaon locally n a sequence of observaons. In such applcaons, he sgnal s eporal and non saonary; he conex of a sngle observaon brngs nforaon abou he evoluon of he sgnal n e. Our purpose n he presen work s o envsage hs approach for sedenal seres deposed durng e. hs work s srucured as follows: Hs are brefly presened n Secon 2, as well as her applcaon o lhology denfcaon. For ha, we sar descrbng he applcave conex n deal. hen, he odel s descrbed n Secon 3, and resuls are hen presened and dscussed n Secon H FOR LIHOLOGY IDENIFICAION 2. he applcave conex: he arcoule se he arcoule se s n he souh of France, n he Gard area, near Bagnols-sur-Cèze. A hundred llon years ago, hs se was covered by an ocean and asde he ounans of he assf Cenral. I s why he subsol s coposed of boh faces of connenal orgn (resulng fro eroson of he crsalln foraons of he assf Cenral and of arne orgn. he subsol s ade of clayey and sandy sedenal seres, whch have been deposed a Creaceous. Daa coe fro wo drllng holes, naed AR402 and AR203. he profle of he arcoule se shows a l of he sol beween hese wo holes; because of hs l, he faces encounered n AR203 are encounered n he nferor half of he well AR402. For ha reason, AR402 s n fac ore coplee han AR203 fro a geologcal pon of vew: soe faces presen n AR402 are absen of AR203. hs s an poran fac n our sudy. Also, core daa fro holes s only avalable a ceran levels of he holes. For hs reason, we use labels resulng fro a prevous research work obaned wh a Kohonen ap on he arcoule se [Frayssne (2000]. AR203 s drlled unl 89 s and AR402 unl 530 s. We use hree logs ha are PEF (phooelecrc effec, RHOB (relave densy n gr/c3, and GR (Gaa-Ray n API nubers. In he drllng, he easures are aken every half-foo (5.24 c. Neverheless, he sudy of sgnals shows ha her vercal resoluon s raher of around 50 cs. In he drllng hole AR203, we have 5590 easures' levels, and n AR402, 9962 easures' levels. Accordng o [Frayssne (2000], welve lhologes were deerned n he arcoule se: Lesones (C, arls (, glauconc Sandsone (Gga, Shales (A, Oher Shales (A, Sls (S, Oher Sls (S, coarse Sandsone (Ggs, Sandsone (G, sandy Lesones (Cg, sandy Brecca (B and Lgne (L. 2.2 Hdden arkov odels In he las years, Hdden arkov odels have becoe a useful ool n non saonary sgnal recognon [Rabner (989, Rabner (993]. Hs are sascal odels based on he classcal arkov chans. Consder a sochasc process (q whch s descrbed a e as beng n one of a se of N saes, S, S 2,,S N. he process (q s a arkov chan f, n order o ake a predcon a e on wha s gong o happen n he

2 fuure, s useless o know anyhng ore abou he whole pas up o e -,.e. ( q s j q s, q 2 sk, q s P( q s j q s P..., ( We only consder he hoogeneous arkov Chan, ha s hose processes n whch he rgh-hand sde of ( (naely he ranson probably fro sae q - o sae q s ndependen of e. he arx of sae ransons probables A{a j } and he nal dsrbuon π are he relevan nforaon n order o descrbe he e evoluon of he process: j ( q s q s a P,,j N (2 P ( q s j π, N (3 he arkov Chan defned n hs way s called an observable arkov odel snce he oupu of he process are he saes. In any neresng probles n whch he sgnal s non saonary, he saes of he arkov Chan are hdden, no drecly observable, and he observaons are he rando sgnals eed by he saes. A Hdden arkov odel (H s herefore a double sochasc process characerzed by: - he nuber N of saes {S, S 2,,S N } n he odel; - he sae ranson probably dsrbuon; a N j j P a j ( q s q s - he nal dsrbuon; ( q s,j N (4 π P, N. (5 - he se of observaon sgnal denses, B{b j }, where b j s he observaon sgnal densy when he process s n sae j. A H provdes he echans for a rando syse whch ay be descrbed as follows. A e, he nal sae q S wll be chosen a rando, accordng o he nal dsrbuon probably π. In hs sae S, a sgnal O wll be observed accordng o he observaon sgnal densy b. A e 2, he process changes o anoher sae S j accordng o he ranson arx a j, and so on. Noe ha a coplee specfcaon of a H s gven by he specfcaon of probably easures A, B and π. In he followng, λ(a,b,π denoes he coplee se of paraeers specfyng he H λ. For a coplee descrpon of ranng procedures n a H, see [Rabner (989, Rabner (993]. o denfy lhologes, we frs ran a H per class (per lhology. Durng hs sep, called ranng, for each lhology, we opze he odels' paraeers (A, B, π ha bes explan a gven se of observaon sequences, called ranng daabase. Aferwards, n a second sep, called recognon, a sequence of logs s, a he sae e, segened and recognzed by he lhologes' Hs. he Verb algorh [Rabner (989] gves ndeed he sequence of saes wh hghes lkelhood for hs observaon sequence. hs allows o segen he observaon sequence correspondng o a log sere of a hole, n dfferen lhologes, whle such lhologes are recognzed. hese wo seps wll be dealed n secon Srucure of he odel 3. RAINING here are dfferen ypes of Hs: dscree or connuous Hs, regardng he naure of he sae esson probably laws, lef-rgh Hs, or parallel ones, or ergodc Hs, regardng o he opology of he odel [Rabner (993], ha s he ransons ha are auhorzed beween he saes of he H. We odel each lhology by an ergodc and gaussan connous H. Ergodcy pers o envsage ransons fro every sae o any oher sae of he H (see Fgure. Also, we used a xure of gaussan denses o approxae he dsrbuon of he observaons, represened by he logs. hs eans ha each observaon O (each log a e, a vecor of denson 3 (he PEF, RHOB and GR logs, s eed by sae j wh probably: b j ( O c [ η O, µ, U ], j N (6 where c s he xure coeffcen for he h xure coponen n sae j and η he gaussan densy funcon, wh ean and covarance arx U for he h µ ( xure coponen n sae j, ha s η O, µ,u : η ( O, µ, U ( 2π n 2 (de( U - 2 exp ( o µ U ( o µ he xure coeffcens c sasfy he sochasc consrans: c c 0 N, (8 s a N a N a 2 a 2 a 2N s N s 2 a N2 Fgure. Ergodc H of a gven lhology (N saes 3.2 Isolaed ranng 2 (7

3 We frs consder a ranng paradg n whch solaed log sequences of each class (lhology are used o ran he correspondng H. We call hs parcular ranng paradg "Isolaed ranng". For hs, we cu he coplee sequence of logs of hole 402 (he hole used for ranng purposes no segens, where each segen corresponds o a dfferen lhology. Each resulng sequence of observaons n a gven lhology has a sze 6. We used he Bau-Welch algorh o esae he paraeers of each lhology H λ(a,b,π. o suarze, hs algorh axzes eravely P(O λ, he lkelhood of he observaon gven he odel. A local axu s aaned afer a gven nuber of eraons of he ranng daabase. hs algorh works n he followng erave for:.- Inalzaon of he odel: he ranson and nal probables are defned equprobable. Also, he nuber of observaons n each log sequence are dsrbued equably n he saes of he H. he sae s done n each sae regardng he nuber of gaussan denses of he sae esson probably law. 2.- Afer each eraon of he ranng daabase, we reesae ˆ λ Aˆ, B ˆ, ˆ as follows: ( Π - for he nal probably, he expeced frequency n sae s a e s copued as: ˆ π ( N (9 where ( s he a poseror probably of beng n sae a e : ( P ( q O λ, (0 - for ranson probables, he expeced nuber of ransons fro sae s o sae s j, dvded by he expeced nuber of ransons fro s s copued: aˆ j where: ε (, j, (,j N ( (, j P ( q, q j O λ ε +, (2 ha s, he a poseror probably o be n sae a e and n sae j a e +; - for he sae esson probables, he paraeers of each gaussan densy funcon and he xure coeffcens are reesaed. he xure coeffcen reesaon for he gaussan densy k n sae j s he followng [Rabner (989]: cˆ k ( j, k ( j, k (3 where ( j, k s he probably of beng n sae j a e wh he kh xure coponen accounng for O ( j, k ( j η [ O, c η µ ] [ O, µ, U ], U (4 Fnally, he ean and he covarance arx of he gaussan densy k n sae j are reesaed as follows: ˆ µ Uˆ ( ( j, k O ( j, k j, k( o µ ( o µ ' ( j, k (5 (6 3.- ranng s sopped when he average log-lkelhood on he ranng daabase s sablzed, ha s when: P r + r ( O P ( O λ r P ( O λ λ < c (7 where c s 0-3 or 0-4 accordng o he class (lhology ha we consder. Bau [Bau (970] showed ha a he rh eraon of hs algorh, we have: P ( O λ r + P r ( O λ for each observaon sequence, unl a local axu s reached. 3.3 Conexual ranng afer Isolaed ranng Afer solaed ranng s perfored, we consder hs as an nalzaon for anoher ype of ranng, ha we call "Conexual ranng".he neres of such ranng s o nroduce n he paraeer esaon process, conexual nforaon presen n he hole. Indeed, solaed ranng only rans each lhology odel on solaed sequences of each lhology. In hs new ranng paradg, we consder nsead longer subsequences of logs of he hole, conanng several lologes. hen, for ranng purposes, we us concaenae he Hs correspondng o he lhologes presen n each of hese subsequences. hs way, paraeer esaon for each of hese lhology odels wll be nfluenced by he neghborng lhologes, as follows : he Verb algorh [Rabner (989] s used o segen he whole sequence no dfferen lhologes, and hs segenaon s exploed for ranng purposes, as we explan below. he Verb algorh copues he opal pah shown n Fgure 2 (he "hdden" sae sequence n ers of axal lkelhood of he whole logs-subsequence (conanng C,, and L, presened o he correspondng

4 Hs (hose of C, and L (see Fgure 2. he opaly creron used on he sequence of logs s herefore global. Conexual nforaon s nroduced n he ranng process precsely when usng he resulng opal pah o reesae he paraeers of he correspondng Hs. Fgure 2. Verb pah copuaon durng Conexual ranng on a logs-sequence alernang C, and L he Verb algorh [Rabner (989] s based on dynac prograng; hs algorh fnds he bes sae sequence n he sense of axal lkelhood, for a gven observaon (logs sequence. For he frs observaons, he probably: ax (... q, OO... O ( q q q Pqq λ (8 gves he bes score along he pah perng a e o reach sae s : A recurrence s hen saed as follows: [ ( a ] b ( O + ( j ax j j + (9 Also, a varable ψ (j conans he bes precedng sae for sae j a e. hs algorh has hree seps:. Inalzaon ( π b (20 ( O Ψ ( 0 2. Recurson: ( j ax [ ( aj ] b j ( o ( j ax [ ( a ] N 2 N (2 ψ arg j j N 3. ernaon P q ax [ ( ] N (22 arg ax [ ( ] he opal sae sequence s obaned by "backrackng", as follows : q Saes of he odels λ L λ λ C ψ ( q + + O..O 5 O 6.. O -3 O Opal pah Logs CCCCLLLLLLL Lhologes -,-2,., (23 As enoned before, afer hs segenaon sep, he reesaon of he Hs' paraeers s perfored. hs ranng paradg s well-known as he Segenal K- eans algorh [Rabner (993]. Accordng o he segenaon gven by he opal sae sequence, all he observaons (logs arbued o a gven sae are affeced o a gven gaussan densy n hs sae. hs s done by he copuaon of he dsance of each observaon o he eans of all he gaussans n hs sae. For each observaon, he gaussan realzng he nu dsance s affeced o such observaon. hen, paraeer reesaon can be perfored for he Hs, as follows: - for he ean of he xure coponen, for a gven sequence of he ranng daabase, ˆ µ ( q ( q j O ( O j ( O c c (24 here s he Kronecker funcon, c denoes cluser (he observaons affeced o he xure coponen, and O denoes he curren observaon a e. - for he covarance arx of xure coponen : Uˆ ( q j ( O c ( O ˆ µ ( ( q j ( O c O ˆ µ (25 he xure coeffcen s reesaed as he nuber of observaons (logs affeced o cluser of sae j dvded by he nuber of observaons affeced o sae j : cˆ ( q j ( O ( q j c 3.4 Conexual ranng afer rando nalzaon (26 Anoher ype of ranng was also esed: consss of Conexual ranng by he Segenal K-eans algorh descrbed n Secon 3.3, when he Hs are no nalzed by Isolaed ranng (prevously descrbed n Secon 3.2. In hs fraework, he ranng daabase (a se of sequences of lengh 25 of AR402 s used once (one epoch o nalze he Hs, usng he "correc pah" nsead of he opal pah copued by he Verb algorh. he "correc pah" s n fac he pah ha we oban when we assocae o each observaon (a log s correc label (he correspondng lhology. Wh hs "correc segenaon" of each sequence of he ranng daabase, we oban a frs esaon of he Hs' paraeers by he Segenal K-eans algorh. 3.5 Convergence crera wo convergence crera are consdered: he frs one s based on he sablzaon of he average log-lkelhood per class; he second one s based on he sablzaon of he perforance of each lhology H (he percenage of correc classfcaon.

5 4. ESING HE SYSE Classfcaon s perfored on he oher hole (AR203 usng he Verb algorh: he coplee sequence of logs easured n AR203 s presened o he 2 lhology Hs a he sae e, o copue he opal sae sequence n he Verb sense. o hs end, ransons are auhorzed fro any sae of any H o any sae of any oher H. 4. esng afer Isolaed ranng he followng able (able shows he nuber of log daa avalable per lhology n AR402 ("Daa" colun n able and he nuber of resulng solaed sequences of each lhology ("oal" colun n able, afer daa n AR402 s cu no segens (of axu lengh 6 of each lhology. Noce ha lhologes S, and Cgs have he fewes nuber of sequences for ranng her respecve Hs. On he oher hand, lhology L has he hghes nuber of sequences for ranng purposes. Class Daa oal Class Daa oal C S Cgs Cga G A Cg A B S L able. Daa descrpon for Isolaed ranng Resuls on AR203 afer Isolaed ranng n sequences of each class exraced fro AR402, are presened n able 2: colun "C" s he class (lhology now nubered fro o 2 (desgnng respecvely C,, Gga, A, A, S, S, Ggs, G, Cg, B, L. Colun "S" s he nuber of saes n he H, colun "" s he nuber of gaussan denses (xure coponens per sae of he H, colun "Daa" gves he nuber of logs per class n AR203, colun "LL" s he nuber of ranng eraons ade, colun "E" s he value of consan c n forula (7 o sop ranng by he Bau-Welch algorh, and colun "%" s he percenage of logs correcly classfed n AR203. Accordng o he nuber of sequences per class, he creron o sop ranng uses a dfferen value of consan c (0-3 or 0-4. Dfferen ess were ade changng he nuber of gaussans (, bu we presen only he resuls wh, as hey are he bes. Resuls are globally good, hey vary fro 44.44% o 96.39%. hs can be explaned by he fac ha here are no enough solaed sequences o ran he Hs n he confguraon n whch esson probables are xures of gaussan denses. Indeed, hs fraework ples uch ore paraeers o esae (several covarance arces, several eans, xure coeffcens. I s why n he followng (Secons 4.2 and 4.3, all he experens are perfored n he fraework of one gaussan densy per sae of he Hs. C S Daa LL E % e e e e e e e e e e e e able 2. es resuls for afer Isolaed ranng 4.2 esng afer Conexual ranng wh Hs nalzed by Isolaed ranng In hs fraework, ranson probables beween dfferen lhology Hs are nroduced n he copuaon of he opal pah by he Verb algorh. hese ranson probables are fxed durng ranng and esng; hey are esaed on he hole AR402 by relave frequences. her role s o favour soe ner-odel ransons, accordng o wha s observed n AR402. For Conexual ranng, we used sequences of lengh 25 of AR402. When ranng s sopped accordng o he perforance creron per H (sablzaon of he percenage of correc classfcaon per class, convergence s reached afer 39 epochs (eraons of he ranng daabase, and afer 5 epochs for he creron of lkelhood sablzaon. able 3 shows resuls on AR203 for boh convergence crera: colun "Q" gves n fac he values n whch he perforance becoes sable n he "ranng hole" (AR402, and colun "%Q" gves he correspondng resuls n he "es hole" (AR203. Analogously, colun "LL" gves he average value of he log-lkelhood per class a convergence (when hs value becoes sable, and colun "%LL" gves he correspondng resuls n he "es hole" (AR203. We frs noce ha he percenage of correc classfcaon s proved n half of he lhologes (classes 2,3,6,7,8, copared o he resuls obaned afer Isolaed ranng.

6 he oher lhologes (classes, 4, 5, 9, 0, 2 for whch resuls are degraded, are very xured n he drllng holes, ha s a sngle or very few observaons (logs of such lhologes are ofen found beween oher lhologes. For hs reason, only few sequences of such lhologes have a sgnfcan lengh durng Conexual ranng. hs s parcularly vsble for lhology 2 (L, for whch pleny of daa are avalable, bu such daa are spread n he drllng holes a os easures' levels. Indeed, hs lhology, of vegeal orgn, s very weak and ends o ge daaged durng he daa acquson process, spreadng self a os easures' levels. In oher words, Conexual ranng s effecve when subsequences of each lhology appearng n he conex of oher classes are of sgnfcan lengh. C S Daa Q %Q LL %LL able 3. es afer Conexual ranng when Hs are nalzed by Isolaed ranng 4.3 esng afer only Conexual ranng In hs fraework, as dealed n Secon 3.4, he ranng daabase s used once (one epoch o nalze he Hs, usng he "correc pah" nsead of he Verb pah. Wh hs "correc segenaon" of he sequences of he ranng daabase, we oban a frs esaon of he Hs' paraeers by he Segenal K-eans algorh. Our goal s o evaluae he nfluence of hs nalzaon, done n a conexual way, when followed by Conexual ranng. Resuls are gven n able eraons (epochs of he ranng daabase were necessary o sop ranng, for boh convergence crera (descrbed n Secon 3.5. In boh cases, resuls are he sae. able 4 shows ha for 2/3 of he classes, he resuls are proved copared o hose presened n able 3. Also, he degradaon of class 2 (L s confred afer hs conexual nalzaon. Copared o Isolaed ranng, wo classes are srongly proved: class 2 ( and class (B, and soe classes lke classes 6 and 7 (S and S, and class 0 (Cg are globally unchanged. hs ay reveal ha he laer are que dffcul o odel. C S Daa Q %Q LL %LL able 4. es afer only Conexual ranng 5. CONCLUSIONS We have proposed an orgnal approach based on Hs o denfy lhologes n a drllng hole. hs sascal approach consders a sequence of logs easured n a drllng hole as a e sere. hs pers o nroduce soe conexual nforaon presen n he sequence of logs when esang he paraeers of he sascal odels of each lhology. A lhology s odeled by a gaussan ergodc H and raned n hree dfferen ways: Isolaed ranng (n whch daa are separaed per lhology for ranng, Conexual ranng afer nalzng he Hs by eans of Isolaed ranng, and only Conexual ranng n whch he Hs are even nalzed n a conexual way. he las paradg proves resuls for 2/3 of he classes relavely o he second one. Soe classes are dffcul o odel n any of such paradgs: class 6 and 7 (S and S, and class 0 (Cg. Also, we noced ha Conexual ranng s uneffecve for hose classes whose sequences are no of sgnfcan lengh when aken n he conex of oher lhologes. For ha reason, only wo classes show he real neres of Conexual ranng relavely o Isolaed ranng: class 2 ( and class (B. On he oher hand, class 2 (L shows anoher l of our approach: hs class appears n he conex of all he ohers because s spread a all he levels of he drllng holes. Resuls for hs class ge degraded wh conexual ranng, and when he nalzaon s also done conexually, hey are even ore degraded. hese prelnary ess show ha, whle n general he nroducon of conex proves he classfcaon

7 accuracy, one has o be careful wh soe classes wh a very changng conex. Our furher work wll explore hs aspec n ore deals and propose soe effcen sraegy o cope wh hs phenoenon. 6. ACKNOWLEDGEENS hs work was parally suppored by he French- Venezuelan Acon ECOS-Nord N V990 (N , CDCH-UCV gran N and Agenda Peroleo Accon odelaje Esocasco Aplcado. 7. REFERENCES Anouar F., Badran F., hra S., 997: Self-Organzed ap, A Probablsc Approach, Proceedngs of he Workshop on Self-Organzed aps, Helsnk Unversy of echnology, Espoo, Fnland, June 4-6. Bau, L H., Pere., Soules G., Wess N., 970: A axzaon echnque ocurrng n he sascal analyss of probablsc funcons of arkov chans, Annals of aheacal Sascs, 4, nº, Busch, J.., Forney, W. G., Berry L. N., 987: Deernaon of lhology fro well logs by sascal analyss, SPE Foraon Evaluaon, vol. 2, Davs, J. C., 986: Sasc and daa analyss n geology, 2 nd ed..new York, John Wley, 273p. Delfner, P., Peyre O., Serra O., 987: Auoac deernaon of lhology fro well logs. SPE Foraon Evaluaon, vol. 2, Doveon J. H., 994: Geologc log analyss usng copuer ehods. A Assoc. Peroleu Geologss, Copuer ehods n Geology, No 2, 69p. Frayssne D., hra S., Badran F., Brqueu L., 2000: Use of Neural Neworks n Log's Daa Processng-Predcon and Rebuldng of lhologc faces. Perophyscs ees Geophyscs, Pars, France, Noveber 6-8. Juang. B. H., L. R. Rabner, 990: he Segenal K- eans algorh for esang paraeers of Hdden arkov odels. IEEE ransacons on Acouscs, Speech and Sgnal Processng, 38, nº 9, Rabner L.R., 989: A uoral on Hdden arkov odels and Seleced Applcaons n Speech Recognon. Proceedngs of he IEEE, 77, No 2, Rabner, L.R., 993: Fundaenals of Speech Recognon. Prence Hall Sgnal Processng Seres, Prence Hall. Rogers. S. J., J. H. Fang, C. L. Karra, and D. A. Sanley, 992: Deernaon of Lhology fro Well Logs usng a Neural Nework. A Assoc. Peroleu Geologss Bullen, 76, No 5,

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

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

Learning Objectives. Self Organization Map. Hamming Distance(1/5) Introduction. Hamming Distance(3/5) Hamming Distance(2/5) 15/04/2015

Learning Objectives. Self Organization Map. Hamming Distance(1/5) Introduction. Hamming Distance(3/5) Hamming Distance(2/5) 15/04/2015 /4/ Learnng Objecves Self Organzaon Map Learnng whou Exaples. Inroducon. MAXNET 3. Cluserng 4. Feaure Map. Self-organzng Feaure Map 6. Concluson 38 Inroducon. Learnng whou exaples. Daa are npu o he syse

More information

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms

Introduction ( Week 1-2) Course introduction A brief introduction to molecular biology A brief introduction to sequence comparison Part I: Algorithms Course organzaon Inroducon Wee -2) Course nroducon A bref nroducon o molecular bology A bref nroducon o sequence comparson Par I: Algorhms for Sequence Analyss Wee 3-8) Chaper -3, Models and heores» Probably

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

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore.

This document is downloaded from DR-NTU, Nanyang Technological University Library, Singapore. Ths documen s downloaded from DR-NTU, Nanyang Technologcal Unversy Lbrary, Sngapore. Tle A smplfed verb machng algorhm for word paron n vsual speech processng( Acceped verson ) Auhor(s) Foo, Say We; Yong,

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

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

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

A Modified Genetic Algorithm Comparable to Quantum GA

A Modified Genetic Algorithm Comparable to Quantum GA A Modfed Genec Algorh Coparable o Quanu GA Tahereh Kahookar Toos Ferdows Unversy of Mashhad _k_oos@wal.u.ac.r Habb Rajab Mashhad Ferdows Unversy of Mashhad h_rajab@ferdows.u.ac.r Absrac: Recenly, researchers

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

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

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

( ) [ ] MAP Decision Rule

( ) [ ] MAP Decision Rule Announcemens Bayes Decson Theory wh Normal Dsrbuons HW0 due oday HW o be assgned soon Proec descrpon posed Bomercs CSE 90 Lecure 4 CSE90, Sprng 04 CSE90, Sprng 04 Key Probables 4 ω class label X feaure

More information

GMM parameter estimation. Xiaoye Lu CMPS290c Final Project

GMM parameter estimation. Xiaoye Lu CMPS290c Final Project GMM paraeer esaon Xaoye Lu M290c Fnal rojec GMM nroducon Gaussan ure Model obnaon of several gaussan coponens Noaon: For each Gaussan dsrbuon:, s he ean and covarance ar. A GMM h ures(coponens): p ( 2π

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

Chapter 6 DETECTION AND ESTIMATION: Model of digital communication system. Fundamental issues in digital communications are

Chapter 6 DETECTION AND ESTIMATION: Model of digital communication system. Fundamental issues in digital communications are Chaper 6 DCIO AD IMAIO: Fndaenal sses n dgal concaons are. Deecon and. saon Deecon heory: I deals wh he desgn and evalaon of decson ang processor ha observes he receved sgnal and gesses whch parclar sybol

More information

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press,

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press, Lecure Sldes for INTRDUCTIN T Machne Learnng ETHEM ALAYDIN The MIT ress, 2004 alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/2ml CHATER 3: Hdden Marov Models Inroducon Modelng dependences n npu; no

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

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

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

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

[ ] 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

INTRODUCTION TO MACHINE LEARNING 3RD EDITION

INTRODUCTION TO MACHINE LEARNING 3RD EDITION ETHEM ALPAYDIN The MIT Press, 2014 Lecure Sdes for INTRODUCTION TO MACHINE LEARNING 3RD EDITION aaydn@boun.edu.r h://www.ce.boun.edu.r/~ehe/23e CHAPTER 7: CLUSTERING Searaerc Densy Esaon 3 Paraerc: Assue

More information

Discrete Markov Process. Introduction. Example: Balls and Urns. Stochastic Automaton. INTRODUCTION TO Machine Learning 3rd Edition

Discrete Markov Process. Introduction. Example: Balls and Urns. Stochastic Automaton. INTRODUCTION TO Machine Learning 3rd Edition EHEM ALPAYDI he MI Press, 04 Lecure Sldes for IRODUCIO O Machne Learnng 3rd Edon alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/ml3e Sldes from exboo resource page. Slghly eded and wh addonal examples

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

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

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

A New Method for Computing EM Algorithm Parameters in Speaker Identification Using Gaussian Mixture Models

A New Method for Computing EM Algorithm Parameters in Speaker Identification Using Gaussian Mixture Models 0 IACSI Hong Kong Conferences IPCSI vol. 9 (0) (0) IACSI Press, Sngaore A New ehod for Comung E Algorhm Parameers n Seaker Idenfcaon Usng Gaussan xure odels ohsen Bazyar +, Ahmad Keshavarz, and Khaoon

More information

Digital Speech Processing Lecture 20. The Hidden Markov Model (HMM)

Digital Speech Processing Lecture 20. The Hidden Markov Model (HMM) Dgal Speech Processng Lecure 20 The Hdden Markov Model (HMM) Lecure Oulne Theory of Markov Models dscree Markov processes hdden Markov processes Soluons o he Three Basc Problems of HMM s compuaon of observaon

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

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation Global Journal of Pure and Appled Mahemacs. ISSN 973-768 Volume 4, Number 6 (8), pp. 89-87 Research Inda Publcaons hp://www.rpublcaon.com Exsence and Unqueness Resuls for Random Impulsve Inegro-Dfferenal

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

Long Term Power Load Combination Forecasting Based on Chaos-Fractal Theory in Beijing

Long Term Power Load Combination Forecasting Based on Chaos-Fractal Theory in Beijing JAGUO ZHOU e al: LOG TERM POWER LOAD COMBIATIO FORECASTIG BASED O CHAOS Long Ter Power Load Cobnaon Forecasng Based on Chaos-Fracal Theory n Bejng Janguo Zhou,We Lu,*,Qang Song School of Econocs and Manageen

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

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

Clustering (Bishop ch 9)

Clustering (Bishop ch 9) Cluserng (Bshop ch 9) Reference: Daa Mnng by Margare Dunham (a slde source) 1 Cluserng Cluserng s unsupervsed learnng, here are no class labels Wan o fnd groups of smlar nsances Ofen use a dsance measure

More information

Consider processes where state transitions are time independent, i.e., System of distinct states,

Consider processes where state transitions are time independent, i.e., System of distinct states, Dgal Speech Processng Lecure 0 he Hdden Marov Model (HMM) Lecure Oulne heory of Marov Models dscree Marov processes hdden Marov processes Soluons o he hree Basc Problems of HMM s compuaon of observaon

More information

Cubic Bezier Homotopy Function for Solving Exponential Equations

Cubic Bezier Homotopy Function for Solving Exponential Equations Penerb Journal of Advanced Research n Compung and Applcaons ISSN (onlne: 46-97 Vol. 4, No.. Pages -8, 6 omoopy Funcon for Solvng Eponenal Equaons S. S. Raml *,,. Mohamad Nor,a, N. S. Saharzan,b and M.

More information

Response of MDOF systems

Response of MDOF systems Response of MDOF syses Degree of freedo DOF: he nu nuber of ndependen coordnaes requred o deerne copleely he posons of all pars of a syse a any nsan of e. wo DOF syses hree DOF syses he noral ode analyss

More information

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University

Hidden Markov Models Following a lecture by Andrew W. Moore Carnegie Mellon University Hdden Markov Models Followng a lecure by Andrew W. Moore Carnege Mellon Unversy www.cs.cmu.edu/~awm/uorals A Markov Sysem Has N saes, called s, s 2.. s N s 2 There are dscree meseps, 0,, s s 3 N 3 0 Hdden

More information

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES

EP2200 Queuing theory and teletraffic systems. 3rd lecture Markov chains Birth-death process - Poisson process. Viktoria Fodor KTH EES EP Queung heory and eleraffc sysems 3rd lecure Marov chans Brh-deah rocess - Posson rocess Vora Fodor KTH EES Oulne for oday Marov rocesses Connuous-me Marov-chans Grah and marx reresenaon Transen and

More information

3D Human Pose Estimation from a Monocular Image Using Model Fitting in Eigenspaces

3D Human Pose Estimation from a Monocular Image Using Model Fitting in Eigenspaces J. Sofware Engneerng & Applcaons, 00, 3, 060-066 do:0.436/jsea.00.35 Publshed Onlne Noveber 00 (hp://www.scrp.org/journal/jsea) 3D Huan Pose Esaon fro a Monocular Iage Usng Model Fng n Egenspaces Gel Bo,

More information

Testing a new idea to solve the P = NP problem with mathematical induction

Testing a new idea to solve the P = NP problem with mathematical induction Tesng a new dea o solve he P = NP problem wh mahemacal nducon Bacground P and NP are wo classes (ses) of languages n Compuer Scence An open problem s wheher P = NP Ths paper ess a new dea o compare he

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

THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Series A, OF THE ROMANIAN ACADEMY Volume 9, Number 1/2008, pp

THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMANIAN ACADEMY, Series A, OF THE ROMANIAN ACADEMY Volume 9, Number 1/2008, pp THE PUBLISHING HOUSE PROCEEDINGS OF THE ROMNIN CDEMY, Seres, OF THE ROMNIN CDEMY Volue 9, Nuber /008, pp. 000 000 ON CIMMINO'S REFLECTION LGORITHM Consann POP Ovdus Unversy of Consana, Roana, E-al: cpopa@unv-ovdus.ro

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

Sampling Procedure of the Sum of two Binary Markov Process Realizations Samplng Procedure of he Sum of wo Bnary Markov Process Realzaons YURY GORITSKIY Dep. of Mahemacal Modelng of Moscow Power Insue (Techncal Unversy), Moscow, RUSSIA, E-mal: gorsky@yandex.ru VLADIMIR KAZAKOV

More information

Modeling of Combined Deterioration of Concrete Structures by Competing Hazard Model

Modeling of Combined Deterioration of Concrete Structures by Competing Hazard Model Modelng of Cobned Deeroraon of Concree Srucures by Copeng Hazard Model Kyoyuk KAITO Assocae Professor Froner Research Cener Osaka Unv., Osaka, apan kao@ga.eng.osaka-u.ac.p Kyoyuk KAITO, born 97, receved

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

Fourier Analysis Models and Their Application to River Flows Prediction

Fourier Analysis Models and Their Application to River Flows Prediction The s Inernaonal Appled Geologcal ongress, Deparen of Geology, Islac Azad Unversy - Mashad Branch, Iran, 6-8 Aprl Fourer Analyss Models and Ther Applcaon o Rver Flows Predcon ohel Ghareagha Zare - Mohaad

More information

A TWO-LEVEL LOAN PORTFOLIO OPTIMIZATION PROBLEM

A TWO-LEVEL LOAN PORTFOLIO OPTIMIZATION PROBLEM Proceedngs of he 2010 Wner Sulaon Conference B. Johansson, S. Jan, J. Monoya-Torres, J. Hugan, and E. Yücesan, eds. A TWO-LEVEL LOAN PORTFOLIO OPTIMIZATION PROBLEM JanQang Hu Jun Tong School of Manageen

More information

Transmit Waveform Selection for Polarimetric MIMO Radar Based on Mutual Information Criterion

Transmit Waveform Selection for Polarimetric MIMO Radar Based on Mutual Information Criterion Sensors & Transducers ol. 5 Specal Issue Deceber 3 pp. 33-38 Sensors & Transducers 3 by IFSA hp://www.sensorsporal.co Trans Wavefor Selecon for Polarerc MIMO Radar Based on Muual Inforaon Creron ajng CUI

More information

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

( 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

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

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon Alernae approach o second order specra: ask abou x magnezaon nsead of energes and ranson probables. If we say wh one bass se, properes vary

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

A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE

A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE S13 A DECOMPOSITION METHOD FOR SOLVING DIFFUSION EQUATIONS VIA LOCAL FRACTIONAL TIME DERIVATIVE by Hossen JAFARI a,b, Haleh TAJADODI c, and Sarah Jane JOHNSTON a a Deparen of Maheacal Scences, Unversy

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

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

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

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

1 Widrow-Hoff Algorithm

1 Widrow-Hoff Algorithm COS 511: heoreical Machine Learning Lecurer: Rob Schapire Lecure # 18 Scribe: Shaoqing Yang April 10, 014 1 Widrow-Hoff Algorih Firs le s review he Widrow-Hoff algorih ha was covered fro las lecure: Algorih

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

A Novel Curiosity-Driven Perception-Action Cognitive Model

A Novel Curiosity-Driven Perception-Action Cognitive Model Inernaonal Conference on Arfcal Inellgence: Technologes and Applcaons (ICAITA 6) A Novel Curosy-Drven Percepon-Acon Cognve Model Jng Chen* Bng L and L L School of Inforaon Technology Engneerng Tanjn Unversy

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

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

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

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov

e-journal Reliability: Theory& Applications No 2 (Vol.2) Vyacheslav Abramov June 7 e-ournal Relably: Theory& Applcaons No (Vol. CONFIDENCE INTERVALS ASSOCIATED WITH PERFORMANCE ANALYSIS OF SYMMETRIC LARGE CLOSED CLIENT/SERVER COMPUTER NETWORKS Absrac Vyacheslav Abramov School

More information

Homework 8: Rigid Body Dynamics Due Friday April 21, 2017

Homework 8: Rigid Body Dynamics Due Friday April 21, 2017 EN40: Dynacs and Vbraons Hoework 8: gd Body Dynacs Due Frday Aprl 1, 017 School of Engneerng Brown Unversy 1. The earh s roaon rae has been esaed o decrease so as o ncrease he lengh of a day a a rae of

More information

Cointegration Analysis of Government R&D Investment and Economic Growth in China

Cointegration Analysis of Government R&D Investment and Economic Growth in China Proceedngs of he 7h Inernaonal Conference on Innovaon & Manageen 349 Conegraon Analyss of Governen R&D Invesen and Econoc Growh n Chna Mao Hu, Lu Fengchao Dalan Unversy of Technology, Dalan,P.R.Chna, 6023

More information

グラフィカルモデルによる推論 確率伝搬法 (2) Kenji Fukumizu The Institute of Statistical Mathematics 計算推論科学概論 II (2010 年度, 後期 )

グラフィカルモデルによる推論 確率伝搬法 (2) Kenji Fukumizu The Institute of Statistical Mathematics 計算推論科学概論 II (2010 年度, 後期 ) グラフィカルモデルによる推論 確率伝搬法 Kenj Fukuzu he Insue of Sascal Maheacs 計算推論科学概論 II 年度 後期 Inference on Hdden Markov Model Inference on Hdden Markov Model Revew: HMM odel : hdden sae fne Inference Coue... for any Naïve

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

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

Speech recognition in noise by using word graph combinations

Speech recognition in noise by using word graph combinations Proceedngs of 0 h Inernaonal Congress on Acouscs, ICA 00 3-7 Augus 00, Sydney, Ausrala Seech recognon n by usng word grah cobnaons Shunsuke Kuraaa, Masaharu Kao and Tesuo Kosaka Graduae School of Scence

More information

Hidden Markov Models

Hidden Markov Models 11-755 Machne Learnng for Sgnal Processng Hdden Markov Models Class 15. 12 Oc 2010 1 Admnsrva HW2 due Tuesday Is everyone on he projecs page? Where are your projec proposals? 2 Recap: Wha s an HMM Probablsc

More information

Changeovers. Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA

Changeovers. Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213, USA wo ew Connuous-e odels for he Schedulng of ulsage Bach Plans wh Sequence Dependen Changeovers Pedro. Casro * gnaco E. Grossann and Auguso Q. ovas Deparaeno de odelação e Sulação de Processos E 649-038

More information

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press,

Lecture Slides for INTRODUCTION TO. Machine Learning. ETHEM ALPAYDIN The MIT Press, Lecure ldes for INRODUCION O Machne Learnng EHEM ALPAYDIN he MI Press, 004 alpaydn@boun.edu.r hp://.cpe.boun.edu.r/~ehe/l CHAPER 6: Densonaly Reducon Why Reduce Densonaly?. Reduces e copley: Less copuaon.

More information

Math 128b Project. Jude Yuen

Math 128b Project. Jude Yuen Mah 8b Proec Jude Yuen . Inroducon Le { Z } be a sequence of observed ndependen vecor varables. If he elemens of Z have a on normal dsrbuon hen { Z } has a mean vecor Z and a varancecovarance marx z. Geomercally

More information

Sklar: Sections (4.4.2 is not covered).

Sklar: Sections (4.4.2 is not covered). COSC 44: Dgal Councaons Insrucor: Dr. Ar Asf Deparen of Copuer Scence and Engneerng York Unversy Handou # 6: Bandpass Modulaon opcs:. Phasor Represenaon. Dgal Modulaon Schees: PSK FSK ASK APK ASK/FSK)

More information

Appendix to Online Clustering with Experts

Appendix to Online Clustering with Experts A Appendx o Onlne Cluserng wh Expers Furher dscusson of expermens. Here we furher dscuss expermenal resuls repored n he paper. Ineresngly, we observe ha OCE (and n parcular Learn- ) racks he bes exper

More information

Density Matrix Description of NMR BCMB/CHEM 8190

Density Matrix Description of NMR BCMB/CHEM 8190 Densy Marx Descrpon of NMR BCMBCHEM 89 Operaors n Marx Noaon If we say wh one bass se, properes vary only because of changes n he coeffcens weghng each bass se funcon x = h< Ix > - hs s how we calculae

More information

Recovering the 3D Shape and Respiratory Motion of the Rib using Chest X-Ray Image

Recovering the 3D Shape and Respiratory Motion of the Rib using Chest X-Ray Image 694 MEDICAL IMAGING ECHNOLOGY Vol.20 No.6 Noveber 2002 Recoverng he 3D Shape and Respraory Moon of he Rb usng Ches X-Ray Iage Myn Myn Sen *, Msuru Kozu *2, Yosho Yanaghara *3 and Hrosu Haa *3 Absrac -hs

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

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

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

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

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

2.1 Constitutive Theory

2.1 Constitutive Theory Secon.. Consuve Theory.. Consuve Equaons Governng Equaons The equaons governng he behavour of maerals are (n he spaal form) dρ v & ρ + ρdv v = + ρ = Conservaon of Mass (..a) d x σ j dv dvσ + b = ρ v& +

More information

Hidden Markov Model for Speech Recognition. Using Modified Forward-Backward Re-estimation Algorithm

Hidden Markov Model for Speech Recognition. Using Modified Forward-Backward Re-estimation Algorithm IJCSI Inernaonal Journal of Compuer Scence Issues Vol. 9 Issue 4 o 2 July 22 ISS (Onlne): 694-84.IJCSI.org 242 Hdden Markov Model for Speech Recognon Usng Modfed Forard-Backard Re-esmaon Algorhm Balan

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

Multi-Perspective Cost-Sensitive Context-Aware Multi-Instance Sparse Coding and Its Application to Sensitive Video Recognition 1

Multi-Perspective Cost-Sensitive Context-Aware Multi-Instance Sparse Coding and Its Application to Sensitive Video Recognition 1 Mul-Perspecve Cos-Sensve Conex-Aware Mul-Insance Sparse Codng and Is Applcaon o Sensve Vdeo Recognon Weng Hu, Xnao Dng, Bng L, Janchao Wang, Yan Gao, Fangsh Wang 3, and Sephen Maybank 4 whu@nlpr.a.ac.cn;

More information

Video-Based Face Recognition Using Adaptive Hidden Markov Models

Video-Based Face Recognition Using Adaptive Hidden Markov Models Vdeo-Based Face Recognon Usng Adapve Hdden Markov Models Xaomng Lu and suhan Chen Elecrcal and Compuer Engneerng, Carnege Mellon Unversy, Psburgh, PA, 523, U.S.A. xaomng@andrew.cmu.edu suhan@cmu.edu Absrac

More information

Influence of Probability of Variation Operator on the Performance of Quantum-Inspired Evolutionary Algorithm for 0/1 Knapsack Problem

Influence of Probability of Variation Operator on the Performance of Quantum-Inspired Evolutionary Algorithm for 0/1 Knapsack Problem The Open Arfcal Inellgence Journal,, 4, 37-48 37 Open Access Influence of Probably of Varaon Operaor on he Perforance of Quanu-Inspred Eoluonary Algorh for / Knapsack Proble Mozael H.A. Khan* Deparen of

More information

Learning for Cognitive Wireless Users

Learning for Cognitive Wireless Users Learnng for Cognve Wreless Users Shor Paper Y Su and Mhaela van der Schaar Dep. of Elecrcal Engneerng UCLA {ysu haela}@ee.ucla.edu Absrac Ths paper sudes he value of learnng for cognve ranscevers n dynac

More information

ECE 366 Honors Section Fall 2009 Project Description

ECE 366 Honors Section Fall 2009 Project Description ECE 366 Honors Secon Fall 2009 Projec Descrpon Inroducon: Muscal genres are caegorcal labels creaed by humans o characerze dfferen ypes of musc. A muscal genre s characerzed by he common characerscs shared

More information

Fitting a transformation: Feature based alignment May 1 st, 2018

Fitting a transformation: Feature based alignment May 1 st, 2018 5//8 Fng a ransforaon: Feaure based algnen Ma s, 8 Yong Jae Lee UC Davs Las e: Deforable conours a.k.a. acve conours, snakes Gven: nal conour (odel) near desred objec Goal: evolve he conour o f eac objec

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL Sco Wsdom, John Hershey 2, Jonahan Le Roux 2, and Shnj Waanabe 2 Deparmen o Elecrcal Engneerng, Unversy o Washngon, Seale, WA, USA

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

Off line signatures Recognition System using Discrete Cosine Transform and VQ/HMM

Off line signatures Recognition System using Discrete Cosine Transform and VQ/HMM Ausralan Journal of Basc and Appled Scences, 6(12): 423-428, 2012 ISSN 1991-8178 Off lne sgnaures Recognon Sysem usng Dscree Cosne Transform and VQ/HMM 1 Behrouz Vasegh, 2 Somayeh Hashem, 1,2 Deparmen

More information