Advanced Machine Learning
|
|
- Franklin Chapman
- 5 years ago
- Views:
Transcription
1 dvaced ache Learg Learg rahcal odels Learg full observed ad arall observed BN rc g Lecure 4 ugus Readg: rc g rc U Iferece ad Learg BN descrbes a uque robabl dsrbuo P Tcal asks: Task : How do we aswer queres abou P? We use ferece as a ae for he rocess of coug aswers o such queres So far we have leared several algorhs for eac ad aro. ferece Task : How do we esae a lausble odel fro daa D?. We use learg as a ae for he rocess of obag o esae of.. Bu for Baesa he seek D whch s acuall a ferece roble.. Whe o all varables are observable eve coug o esae of eed o do ferece o ue he ssg daa. rc g rc U
2 Learg rahcal odels The goal: ve se of deede sales assges of rado varables fd he bes he os lkel? grahcal odel boh he grah ad he PDs B B R R BRTFFTF BRTFTTF.. BRFTTTF rc g rc U e e e e B b b b b P B Learg rahcal odels Scearos: coleel observed s dreced udreced arall observed s dreced udreced a oe research oc sao rcles: aal lkelhood esao L Baesa esao aal codoal lkelhood aal "arg" We use learg as a ae for he rocess of esag he araeers ad soe cases he oo of he ework fro daa. rc g rc U
3 Score-based aroach Daa Possble srucures B Lear araeers Score sruc/ara R au lkelhood 0 5 K K K K R B Baesa odoal lkelhood arg K. rc g rc U Z L Paraeer s. for coleel observed s of gve srucure The daa: { N N } rc g rc U
4 4 rc g rc U Lkelhood for ow le's assue ha he srucure s gve: Log-Lkelhood: Daa -lkelhood L 3 ; L l DT l 3 3 arg a } { DT l L 3 arg a arg a a arg * * * The basc dea uderlg L rc g rc U The coleel observed odel: Z s a class dcaor vecor s a codoal aussa varable wh a class-secfc ea Z 0 ad ] [ where Z Z Z Z Z Z K ll ece oe of hese ers wll be oe { } - e - / µ σ σ µ σ N σ µ σ µ ad a dau s class w.. ale : codoal aussa
5 ale : codoal aussa Daa -lkelhood L l D µ σ µ σ * N σ - µ + * * µ arg a l D µ he average of rc g rc U µ σ * arg a l D l D 0 s.. N N Z he fraco of sales of class sales of class ale : H: wo scearos Suervsed learg: esao whe he rgh aswer s kow ales: IVN: IVN: a geoc rego where we have good eereal aoaos of he slads he caso laer allows us o observe h oe eveg as he chages dce ad roduces 0000 rolls Usuervsed learg: esao whe he rgh aswer s ukow ales: IVN: IVN: he orcue geoe; we do kow how freque are he slads here eher do we kow her cooso 0000 rolls of he caso laer bu we do see whe he chages dce QUSTION: Udae he araeers of he odel o ae P --- aal lkelhood L esao rc g rc U
6 Recall defo of H Traso robables bewee a wo saes 3... T a 3... T or a a K a. ~ uloal I Sar robables ~ uloal K. sso robables assocaed wh each sae b b K b. ~ uloal K I or geeral:. ~ f I rc g rc U Suervsed L esao ve N for whch he rue sae ah N s kow Defe: B k # es sae raso occurs # es sae es k We ca show ha he au lkelhood araeers are: a b L L k # # # k # T T : : T : N Wha f s couous? We ca rea as N T observaos of e.g. a aussa ad al learg rules for aussa ' ' T k Bk T B k ' k ' { } rc g rc U
7 Suervsed L esao cd. Iuo: Whe we kow he uderlg saes he bes esae of s he average frequec of rasos & essos ha occur he rag daa Drawback: ve lle daa here a be overfg: P s aed bu s ureasoable 0 robables VRY BD ale: ve 0 caso rolls we observe F F F F F F F F F F The: a FF ; a FL 0 b F b F3.; b F.3; b F4 0; b F5 b F6. rc g rc U Pseudocous Soluo for sall rag ses: dd seudocous B k # es sae raso occurs + R # es sae es k + S k R S are seudocous rereseg our ror belef Toal seudocous: R Σ R S Σ k S k --- "sregh" of ror belef --- oal uber of agar saces he ror Larger oal seudocous srog ror belef Sall oal seudocous: us o avod 0 robables --- soohg Ths s equvale o Baesa es. uder a ufor ror wh "araeer sregh" equals o he seudocous rc g rc U
8 L for geeral BN araeers If we assue he araeers for each PD are globall deede ad all odes are full observed he he lkelhood fuco decooses o a su of local ers oe er ode: l ; D D rc g rc U ale: decoosable lkelhood of a dreced odel osder he dsrbuo defed b he dreced acclc : Ths s eacl lke learg four searae sall BNs each of whch cosss of a ode ad s ares rc g rc U
9 .g.: L for BNs wh abular PDs ssue each PD s rereseed as a able uloal where def k Noe ha case of ulle ares wll have a coose sae ad he PD wll be a hgh-desoal able The suffce sascs are cous of fal cofguraos The -lkelhood s Usg a Lagrage uller o eforce we ge: k k k def k k l ; D k k k L k ' k k k ' k k rc g rc U Z Learg arall observed s The daa: { 3... N } rc g rc U
10 Wha f soe odes are o observed? osder he dsrbuo defed b he dreced acclc : Need o coue H V ferece rc g rc U Recall: lgorh wa of ag lkelhood fuco for lae varable odels. Fds L of araeers whe he orgal hard roble ca be broke u o wo eas eces:. sae soe ssg or uobserved daa fro observed daa ad curre araeers.. Usg hs colee daa fd he au lkelhood araeer esaes. lerae bewee fllg he lae varables usg he bes guess oseror ad udag he araeers based o hs guess: -se: -se: q + arg af q q + + arg af q I he -se we oe a lower boud o he lkelhood. I he - se we close he ga akg boudlkelhood. rc g rc U
11 for geeral BNs whle o coverged % -se for each ode SS 0 % rese eeced suffce sascs for each daa sale do ferece wh H for each ode SS SS % -se for each ode : LSS + H H rc g rc U ale: H Suervsed learg: esao whe he rgh aswer s kow ales: IVN: a geoc rego where we have good eereal aoaos of he slads IVN: he caso laer allows us o observe h oe eveg as he chages dce ad roduces 0000 rolls Usuervsed learg: esao whe he rgh aswer s ukow ales: IVN: IVN: he orcue geoe; we do kow how freque are he slads here eher do we kow her cooso 0000 rolls of he caso laer bu we do see whe he chages dce QUSTION: Udae he araeers of he odel o ae P - -- aal lkelhood L esao rc g rc U
12 rc g rc U The Bau Welch algorh The colee lkelhood The eeced colee lkelhood The se The se "sbolcall" decal o L T T c ; l + + T k k T c b a ; l γ ξ T T L a γ ξ T k T L k b γ γ N L γ rc g rc U Usuervsed L esao ve N for whch he rue sae ah N s ukow PTTION IIZTION 0. Sarg wh our bes guess of a odel araeers :. sae B k he rag daa How?. Udae accordg o B k Now a "suervsed learg" roble 3. Reea & ul covergece Ths s called he Bau-Welch lgorh We ca ge o a rovabl ore or equall lkel araeer se each erao k k B
13 3 rc g rc U L Srucural Learg for coleel observed s R B Daa K K K K rc g rc U Iforao Theorec Ierreao of L cou D D ˆ ; l Fro su over daa os o su over cou of varable saes
14 4 rc g rc U Iforao Theorec Ierreao of L co'd H I D D ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ˆ ; l Decoosable score ad a fuco of he grah srucure rc g rc U Srucural Search How a grahs over odes? How a rees over odes? Bu urs ou ha we ca fd eac soluo of a oal ree uder L! Trck: a ree each ode has ol oe are! how-lu algorh O O!
15 how-lu ree learg algorh Obeco fuco: l ; D ˆ D how-lu: Iˆ Hˆ Iˆ For each ar of varable ad oue ercal dsrbuo: oue uual forao: cou ˆ Iˆ ˆ ˆ ˆ ˆ Defe a grah wh ode dge I ges wegh Iˆ rc g rc U how-lu algorh co'd Obeco fuco: l ; D ˆ D Iˆ Hˆ Iˆ how-lu: Oal ree BN oue au wegh sag ree Dreco BN: ck a ode as roo do breadh-frs-search o defe drecos I-equvalece: D B D B D B I B + I + I D + I rc g rc U
16 Srucure Learg for geeral grahs Theore: The roble of learg a BN srucure wh a os d ares s NP-hard for a fed d os srucure learg aroaches use heurscs lo score decooso Two heurscs ha elo decooso dffere was rc g reed search hrough sace of ode-orders Local search of grah srucures rc U ee resso Proflg b croarras Receor Kase Receor B 3 Kase D Kase 4 Tras. Facor 6 F ee rc g 7 rc U ee H 5 F F 8 3 6
17 croarra Daa hr hr 3hr 4hr rc g rc U Srucure Learg lgorhs Srucural Freda 998 The orgal algorh resso daa Learg lgorh B R Sarse addae lgorh Freda e al. Dscreg arra sgals Hll-clbg search usg local oeraors: add/delee/swa of a sgle edge Feaure eraco: arkov relaos order relaos Re-asseble hgh-cofdece sub-eworks fro feaures odule ework learg Segal e al. Heursc search of srucure a "odule grah" odule assge Paraeer sharg Pror kowledge: ossble regulaors TF gees rc g rc U
18 Scorg Neworks B D Lear R resale B D resale D Lear R resale... B D Lear R rc g rc U Learg srucure Learg of bes PDs gve D s eas collec sascs of values of each ode gve secfc assge o s ares Learg of he grah oo srucure s NP-hard heursc search us be aled geerall leads o a locall oal ework Overfg I urs ou ha rcher srucures gve hgher lkelhood PD o he daa addg a edge s alwas referable B B P P B ore araeers o f > ore freedo > alwas es ore "oal" PD We refer sler ore elaaor eworks Praccal scores regulare he lkelhood rovee cole eworks. rc g rc U
19 Learg sarse ulvarae aussa over all couous eressos [... ] Σ The recso ar KΣ reveals he oo of he udreced ework K / K dge ~ K > 0 - e - T - { } r - µ Σ r - µ Learg lgorh: ovarace seleco Wa a sarse ar Regresso for each ode wh degree cosra Dobra e al. Regresso for each ode wh herarchcal Baesa ror L e al rahcal Lasso we wll descrbe shorl rc g rc U Learg Isg odel.e. arwse RF ssug he odes are dscree ad edges are weghed he for a sale d we have rah lasso has bee used o oba a sarse esae of wh couous We ca use grahcal L_ regulared sc regresso o oba a sarse esae of wh dscree rc g rc U
20 Recall lasso rc g rc U rah Regresso Lasso: rc g rc U
21 rah Regresso rc g rc U rah Regresso rc g rc U
22 ossec Theore: for he grahcal regresso algorh uder cera verfable codos oed here for slc: Noe he fro hs heore oe should see ha he regularer s o acuall used o roduce a arfcal sars bas bu a devse o esure cossec uder fe daa ad hgh deso codo. rc g rc U Learg Learg of bes PDs gve D s eas collec sascs of values of each ode gve secfc assge o s ares Learg of he grah oo srucure s NP-hard heursc search us be aled geerall leads o a locall oal ework We refer sler ore elaaor eworks Regulared grah regresso rc g rc U
Machine Learning. Hidden Markov Model. Eric Xing / /15-781, 781, Fall Lecture 17, March 24, 2008
Mache Learg 0-70/5 70/5-78 78 Fall 2008 Hdde Marov Model Erc Xg Lecure 7 March 24 2008 Readg: Cha. 3 C.B boo Erc Xg Erc Xg 2 Hdde Marov Model: from sac o damc mure models Sac mure Damc mure Y Y Y 2 Y 3
More informationThree Main Questions on HMMs
Mache Learg 0-70/5-78 78 Srg 00 Hdde Marov Model II Erc Xg Lecure Februar 4 00 Readg: Cha. 3 CB Three Ma Quesos o HMMs. Evaluao GIVEN a HMM M ad a sequece FIND Prob M ALGO. Forward. Decodg GIVEN a HMM
More informationDensity estimation. Density estimations. CS 2750 Machine Learning. Lecture 5. Milos Hauskrecht 5329 Sennott Square
Lecure 5 esy esmao Mlos Hauskrec mlos@cs..edu 539 Seo Square esy esmaos ocs: esy esmao: Mamum lkelood ML Bayesa arameer esmaes M Beroull dsrbuo. Bomal dsrbuo Mulomal dsrbuo Normal dsrbuo Eoeal famly Noaramerc
More information8. Queueing systems lect08.ppt S Introduction to Teletraffic Theory - Fall
8. Queueg sysems lec8. S-38.45 - Iroduco o Teleraffc Theory - Fall 8. Queueg sysems Coes Refresher: Smle eleraffc model M/M/ server wag laces M/M/ servers wag laces 8. Queueg sysems Smle eleraffc model
More informationLeast Squares Fitting (LSQF) with a complicated function Theexampleswehavelookedatsofarhavebeenlinearintheparameters
Leas Squares Fg LSQF wh a complcaed fuco Theeampleswehavelookedasofarhavebeelearheparameers ha we have bee rg o deerme e.g. slope, ercep. For he case where he fuco s lear he parameers we ca fd a aalc soluo
More informationThe ray paths and travel times for multiple layers can be computed using ray-tracing, as demonstrated in Lab 3.
C. Trael me cures for mulple reflecors The ray pahs ad rael mes for mulple layers ca be compued usg ray-racg, as demosraed Lab. MATLAB scrp reflec_layers_.m performs smple ray racg. (m) ref(ms) ref(ms)
More informationDensity estimation III.
Lecure 4 esy esmao III. Mlos Hauskrec mlos@cs..edu 539 Seo Square Oule Oule: esy esmao: Mamum lkelood ML Bayesa arameer esmaes MP Beroull dsrbuo. Bomal dsrbuo Mulomal dsrbuo Normal dsrbuo Eoeal famly Eoeal
More informationThe Poisson Process Properties of the Poisson Process
Posso Processes Summary The Posso Process Properes of he Posso Process Ierarrval mes Memoryless propery ad he resdual lfeme paradox Superposo of Posso processes Radom seleco of Posso Pos Bulk Arrvals ad
More informationLinear Regression Linear Regression with Shrinkage
Lear Regresso Lear Regresso h Shrkage Iroduco Regresso meas predcg a couous (usuall scalar oupu from a vecor of couous pus (feaures x. Example: Predcg vehcle fuel effcec (mpg from 8 arbues: Lear Regresso
More information14. Poisson Processes
4. Posso Processes I Lecure 4 we roduced Posso arrvals as he lmg behavor of Bomal radom varables. Refer o Posso approxmao of Bomal radom varables. From he dscusso here see 4-6-4-8 Lecure 4 " arrvals occur
More informationReal-Time Systems. Example: scheduling using EDF. Feasibility analysis for EDF. Example: scheduling using EDF
EDA/DIT6 Real-Tme Sysems, Chalmers/GU, 0/0 ecure # Updaed February, 0 Real-Tme Sysems Specfcao Problem: Assume a sysem wh asks accordg o he fgure below The mg properes of he asks are gve he able Ivesgae
More informationMidterm Exam. Tuesday, September hour, 15 minutes
Ecoomcs of Growh, ECON560 Sa Fracsco Sae Uvers Mchael Bar Fall 203 Mderm Exam Tuesda, Sepember 24 hour, 5 mues Name: Isrucos. Ths s closed boo, closed oes exam. 2. No calculaors of a d are allowed. 3.
More information(1) Cov(, ) E[( E( ))( E( ))]
Impac of Auocorrelao o OLS Esmaes ECON 3033/Evas Cosder a smple bvarae me-seres model of he form: y 0 x The four key assumpos abou ε hs model are ) E(ε ) = E[ε x ]=0 ) Var(ε ) =Var(ε x ) = ) Cov(ε, ε )
More informationChapter 3: Maximum-Likelihood & Bayesian Parameter Estimation (part 1)
Aoucemes Reags o E-reserves Proec roosal ue oay Parameer Esmao Bomercs CSE 9-a Lecure 6 CSE9a Fall 6 CSE9a Fall 6 Paer Classfcao Chaer 3: Mamum-Lelhoo & Bayesa Parameer Esmao ar All maerals hese sles were
More informationContinuous Time Markov Chains
Couous me Markov chas have seay sae probably soluos f a oly f hey are ergoc, us lke scree me Markov chas. Fg he seay sae probably vecor for a couous me Markov cha s o more ffcul ha s he scree me case,
More informationFundamentals of Speech Recognition Suggested Project The Hidden Markov Model
. Projec Iroduco Fudameals of Speech Recogo Suggesed Projec The Hdde Markov Model For hs projec, s proposed ha you desg ad mpleme a hdde Markov model (HMM) ha opmally maches he behavor of a se of rag sequeces
More informationQR factorization. Let P 1, P 2, P n-1, be matrices such that Pn 1Pn 2... PPA
QR facorzao Ay x real marx ca be wre as AQR, where Q s orhogoal ad R s upper ragular. To oba Q ad R, we use he Householder rasformao as follows: Le P, P, P -, be marces such ha P P... PPA ( R s upper ragular.
More informationCS344: Introduction to Artificial Intelligence
C344: Iroduco o Arfcal Iellgece Puhpa Bhaacharyya CE Dep. IIT Bombay Lecure 3 3 32 33: Forward ad bacward; Baum elch 9 h ad 2 March ad 2 d Aprl 203 Lecure 27 28 29 were o EM; dae 2 h March o 8 h March
More informationθ = θ Π Π Parametric counting process models θ θ θ Log-likelihood: Consider counting processes: Score functions:
Paramerc coug process models Cosder coug processes: N,,..., ha cou he occurreces of a eve of eres for dvduals Iesy processes: Lelhood λ ( ;,,..., N { } λ < Log-lelhood: l( log L( Score fucos: U ( l( log
More informationSolution of Impulsive Differential Equations with Boundary Conditions in Terms of Integral Equations
Joural of aheacs ad copuer Scece (4 39-38 Soluo of Ipulsve Dffereal Equaos wh Boudary Codos Ters of Iegral Equaos Arcle hsory: Receved Ocober 3 Acceped February 4 Avalable ole July 4 ohse Rabba Depare
More informationSpeech, NLP and the Web
peech NL ad he Web uhpak Bhaacharyya CE Dep. IIT Bombay Lecure 38: Uuperved learg HMM CFG; Baum Welch lecure 37 wa o cogve NL by Abh Mhra Baum Welch uhpak Bhaacharyya roblem HMM arg emac ar of peech Taggg
More informationLearning of Graphical Models Parameter Estimation and Structure Learning
Learg of Grahal Models Parameer Esmao ad Sruure Learg e Fukumzu he Isue of Sasal Mahemas Comuaoal Mehodology Sasal Iferee II Work wh Grahal Models Deermg sruure Sruure gve by modelg d e.g. Mxure model
More informationThe Linear Regression Of Weighted Segments
The Lear Regresso Of Weghed Segmes George Dael Maeescu Absrac. We proposed a regresso model where he depede varable s made o up of pos bu segmes. Ths suao correspods o he markes hroughou he da are observed
More informationSolution set Stat 471/Spring 06. Homework 2
oluo se a 47/prg 06 Homework a Whe he upper ragular elemes are suppressed due o smmer b Le Y Y Y Y A weep o he frs colum o oba: A ˆ b chagg he oao eg ad ec YY weep o he secod colum o oba: Aˆ YY weep o
More informationChapter 8. Simple Linear Regression
Chaper 8. Smple Lear Regresso Regresso aalyss: regresso aalyss s a sascal mehodology o esmae he relaoshp of a respose varable o a se of predcor varable. whe here s jus oe predcor varable, we wll use smple
More informationDensity estimation III.
Lecure 6 esy esmao III. Mlos Hausrec mlos@cs..eu 539 Seo Square Oule Oule: esy esmao: Bomal srbuo Mulomal srbuo ormal srbuo Eoeal famly aa: esy esmao {.. } a vecor of arbue values Objecve: ry o esmae e
More informationNUMERICAL EVALUATION of DYNAMIC RESPONSE
NUMERICAL EVALUATION of YNAMIC RESPONSE Aalycal solo of he eqao of oo for a sgle degree of freedo syse s sally o ossble f he excao aled force or grod accelerao ü g -vares arbrarly h e or f he syse s olear.
More informationCyclone. Anti-cyclone
Adveco Cycloe A-cycloe Lorez (963) Low dmesoal aracors. Uclear f hey are a good aalogy o he rue clmae sysem, bu hey have some appealg characerscs. Dscusso Is he al codo balaced? Is here a al adjusme
More informationFor the plane motion of a rigid body, an additional equation is needed to specify the state of rotation of the body.
The kecs of rgd bodes reas he relaoshps bewee he exeral forces acg o a body ad he correspodg raslaoal ad roaoal moos of he body. he kecs of he parcle, we foud ha wo force equaos of moo were requred o defe
More informationAxiomatic Definition of Probability. Problems: Relative Frequency. Event. Sample Space Examples
Rado Sgals robabl & Rado Varables: Revew M. Sa Fadal roessor o lecrcal geerg Uvers o evada Reo Soe phscal sgals ose cao be epressed as a eplc aheacal orla. These sgals s be descrbed probablsc ers. ose
More informationReal-time Classification of Large Data Sets using Binary Knapsack
Real-me Classfcao of Large Daa Ses usg Bary Kapsack Reao Bru bru@ds.uroma. Uversy of Roma La Sapeza AIRO 004-35h ANNUAL CONFERENCE OF THE ITALIAN OPERATIONS RESEARCH Sepember 7-0, 004, Lecce, Ialy Oule
More informationComparison of the Bayesian and Maximum Likelihood Estimation for Weibull Distribution
Joural of Mahemacs ad Sascs 6 (2): 1-14, 21 ISSN 1549-3644 21 Scece Publcaos Comarso of he Bayesa ad Maxmum Lkelhood Esmao for Webull Dsrbuo Al Omar Mohammed Ahmed, Hadeel Salm Al-Kuub ad Noor Akma Ibrahm
More informationLeast squares and motion. Nuno Vasconcelos ECE Department, UCSD
Leas squares ad moo uo Vascocelos ECE Deparme UCSD Pla for oda oda we wll dscuss moo esmao hs s eresg wo was moo s ver useful as a cue for recogo segmeao compresso ec. s a grea eample of leas squares problem
More information4. THE DENSITY MATRIX
4. THE DENSTY MATRX The desy marx or desy operaor s a alerae represeao of he sae of a quaum sysem for whch we have prevously used he wavefuco. Alhough descrbg a quaum sysem wh he desy marx s equvale o
More informationMarch 14, Title: Change of Measures for Frequency and Severity. Farrokh Guiahi, Ph.D., FCAS, ASA
March 4, 009 Tle: Chage of Measures for Frequecy ad Severy Farroh Guah, Ph.D., FCAS, ASA Assocae Professor Deare of IT/QM Zarb School of Busess Hofsra Uversy Hesead, Y 549 Eal: Farroh.Guah@hofsra.edu Phoe:
More informationEE 6885 Statistical Pattern Recognition
EE 6885 Sascal Paer Recogo Fall 005 Prof. Shh-Fu Chag hp://www.ee.columba.edu/~sfchag Lecure 5 (9//05 4- Readg Model Parameer Esmao ML Esmao, Chap. 3. Mure of Gaussa ad EM Referece Boo, HTF Chap. 8.5 Teboo,
More informationContent. A Strange World. Clustering. Introduction. Unsupervised Learning Networks. What is Unsupervised Learning? Unsupervised Learning Networks
Usupervsed Learg Newors Cluserg Coe Iroduco Ipora Usupervsed Learg NNs Hag Newors Kohoe s Self-Orgazg Feaure Maps Grossberg s AR Newors Couerpropagao Newors Adapve BAN Neocogro Cocluso Usupervsed Learg
More informationELEC 6041 LECTURE NOTES WEEK 3 Dr. Amir G. Aghdam Concordia University
ecre Noe Prepared b r G. ghda EE 64 ETUE NTE WEE r. r G. ghda ocorda Uer eceraled orol e - Whe corol heor appled o a e ha co of geographcall eparaed copoe or a e cog of a large ber of p-op ao ofe dered
More informationINTERIOR POINT ALGORITHMS FOR NONLINEAR CONSTRAINED LEAST SQUARES PROBLEMS
4 h Ieraoal Coferece o Iverse Probles Egeerg Ro de Jaero, Brazl, INTERIOR POINT ALGORITHMS FOR NONLINEAR CONSTRAINED LEAST SQUARES PROBLEMS José Hersovs*, Verase Dubeu* *Mechacal Egeerg Progra, COPPE Federal
More informationCS 2750 Machine Learning. Lecture 7. Linear regression. CS 2750 Machine Learning. Linear regression. is a linear combination of input components x
CS 75 Mache Learg Lecture 7 Lear regresso Mlos Hauskrecht los@cs.ptt.edu 59 Seott Square CS 75 Mache Learg Lear regresso Fucto f : X Y s a lear cobato of put copoets f + + + K d d K k - paraeters eghts
More informationFinal Exam Applied Econometrics
Fal Eam Appled Ecoomercs. 0 Sppose we have he followg regresso resl: Depede Varable: SAT Sample: 437 Iclded observaos: 437 Whe heeroskedasc-cosse sadard errors & covarace Varable Coeffce Sd. Error -Sasc
More informationInterval Estimation. Consider a random variable X with a mean of X. Let X be distributed as X X
ECON 37: Ecoomercs Hypohess Tesg Iervl Esmo Wh we hve doe so fr s o udersd how we c ob esmors of ecoomcs reloshp we wsh o sudy. The queso s how comforble re we wh our esmors? We frs exme how o produce
More informationDetermination of Antoine Equation Parameters. December 4, 2012 PreFEED Corporation Yoshio Kumagae. Introduction
refeed Soluos for R&D o Desg Deermao of oe Equao arameers Soluos for R&D o Desg December 4, 0 refeed orporao Yosho Kumagae refeed Iroduco hyscal propery daa s exremely mpora for performg process desg ad
More informationPartial Molar Properties of solutions
Paral Molar Properes of soluos A soluo s a homogeeous mxure; ha s, a soluo s a oephase sysem wh more ha oe compoe. A homogeeous mxures of wo or more compoes he gas, lqud or sold phase The properes of a
More informationFault Tolerant Computing. Fault Tolerant Computing CS 530 Probabilistic methods: overview
Probably 1/19/ CS 53 Probablsc mehods: overvew Yashwa K. Malaya Colorado Sae Uversy 1 Probablsc Mehods: Overvew Cocree umbers presece of uceray Probably Dsjo eves Sascal depedece Radom varables ad dsrbuos
More informationModeling and Predicting Sequences: HMM and (may be) CRF. Amr Ahmed Feb 25
Modelg d redcg Sequeces: HMM d m be CRF Amr Ahmed 070 Feb 25 Bg cure redcg Sgle Lbel Ipu : A se of feures: - Bg of words docume - Oupu : Clss lbel - Topc of he docume - redcg Sequece of Lbels Noo Noe:
More informationOutline. Computer Networks: Theory, Modeling, and Analysis. Delay Models. Queuing Theory Framework. Delay Models. Little s Theorem
Oule Couer Newors: Theory, Modelg, ad Aalyss Guevara Noubr COM35, lecure 3 Delay Models Lle s Theore The M/M/ queug syse The M/G/ queug syse F, COM35 Couer Newors Lecure 3, F, COM35 Couer Newors Lecure
More informationFALL HOMEWORK NO. 6 - SOLUTION Problem 1.: Use the Storage-Indication Method to route the Input hydrograph tabulated below.
Jorge A. Ramírez HOMEWORK NO. 6 - SOLUTION Problem 1.: Use he Sorage-Idcao Mehod o roue he Ipu hydrograph abulaed below. Tme (h) Ipu Hydrograph (m 3 /s) Tme (h) Ipu Hydrograph (m 3 /s) 0 0 90 450 6 50
More informationSolution. The straightforward approach is surprisingly difficult because one has to be careful about the limits.
ose ad Varably Homewor # (8), aswers Q: Power spera of some smple oses A Posso ose A Posso ose () s a sequee of dela-fuo pulses, eah ourrg depedely, a some rae r (More formally, s a sum of pulses of wdh
More informationLaplace Transform. Definition of Laplace Transform: f(t) that satisfies The Laplace transform of f(t) is defined as.
Lplce Trfor The Lplce Trfor oe of he hecl ool for olvg ordry ler dfferel equo. - The hoogeeou equo d he prculr Iegrl re olved oe opero. - The Lplce rfor cover he ODE o lgerc eq. σ j ple do. I he pole o
More informationEfficient Estimators for Population Variance using Auxiliary Information
Global Joural of Mahemacal cece: Theor ad Praccal. IN 97-3 Volume 3, Number (), pp. 39-37 Ieraoal Reearch Publcao Houe hp://www.rphoue.com Effce Emaor for Populao Varace ug Aular Iformao ubhah Kumar Yadav
More informationInterval Regression Analysis with Reduced Support Vector Machine
Ieraoal DSI / Asa ad Pacfc DSI 007 Full Paper (July, 007) Ierval Regresso Aalyss wh Reduced Suppor Vecor Mache Cha-Hu Huag,), Ha-Yg ao ) ) Isue of Iforao Maagee, Naoal Chao Tug Uversy (leohkko@yahoo.co.w)
More informationAs evident from the full-sample-model, we continue to assume that individual errors are identically and
Maxmum Lkelhood smao Greee Ch.4; App. R scrp modsa, modsb If we feel safe makg assumpos o he sascal dsrbuo of he error erm, Maxmum Lkelhood smao (ML) s a aracve alerave o Leas Squares for lear regresso
More informationAML710 CAD LECTURE 12 CUBIC SPLINE CURVES. Cubic Splines Matrix formulation Normalised cubic splines Alternate end conditions Parabolic blending
CUIC SLINE CURVES Cubc Sples Marx formulao Normalsed cubc sples Alerae ed codos arabolc bledg AML7 CAD LECTURE CUIC SLINE The ame sple comes from he physcal srume sple drafsme use o produce curves A geeral
More informationQuantum Mechanics II Lecture 11 Time-dependent perturbation theory. Time-dependent perturbation theory (degenerate or non-degenerate starting state)
Pro. O. B. Wrgh, Auum Quaum Mechacs II Lecure Tme-depede perurbao heory Tme-depede perurbao heory (degeerae or o-degeerae sarg sae) Cosder a sgle parcle whch, s uperurbed codo wh Hamloa H, ca exs a superposo
More informationStrong Convergence Rates of Wavelet Estimators in Semiparametric Regression Models with Censored Data*
8 The Ope ppled Maheacs Joural 008 8-3 Srog Covergece Raes of Wavele Esaors Separaerc Regresso Models wh Cesored Daa Hogchag Hu School of Maheacs ad Sascs Hube Noral Uversy Huagsh 43500 Cha bsrac: The
More informationThe Properties of Probability of Normal Chain
I. J. Coep. Mah. Sceces Vol. 8 23 o. 9 433-439 HIKARI Ld www.-hkar.co The Properes of Proaly of Noral Cha L Che School of Maheacs ad Sascs Zheghou Noral Uversy Zheghou Cy Hea Provce 4544 Cha cluu6697@sa.co
More informationFuzzy Possibility Clustering Algorithm Based on Complete Mahalanobis Distances
Ieraoal Joural of Sef Egeerg ad See Volue Issue. 38-43 7. ISSN (Ole): 456-736 Fuzzy Possbly Cluserg Algorh Based o Colee ahalaobs Dsaes Sue-Fe Huag Deare of Dgal Gae ad Aao Desg Tae Uvey of are Tehology
More informationThe conditional density p(x s ) Bayes rule explained. Bayes rule for a classification problem INF
INF 4300 04 Mulvarae clafcao Ae Solberg ae@fuoo Baed o Chaper -6 Duda ad Har: Paer Clafcao Baye rule for a clafcao proble Suppoe we have J, =,J clae he cla label for a pel, ad he oberved feaure vecor We
More informationOutline. Queuing Theory Framework. Delay Models. Fundamentals of Computer Networking: Introduction to Queuing Theory. Delay Models.
Oule Fudaeals of Couer Neworg: Iroduco o ueug Theory eadg: Texboo chaer 3. Guevara Noubr CSG5, lecure 3 Delay Models Lle s Theore The M/M/ queug syse The M/G/ queug syse F3, CSG5 Fudaeals of Couer Neworg
More informationPGE 310: Formulation and Solution in Geosystems Engineering. Dr. Balhoff. Interpolation
PGE 30: Formulato ad Soluto Geosystems Egeerg Dr. Balhoff Iterpolato Numercal Methods wth MATLAB, Recktewald, Chapter 0 ad Numercal Methods for Egeers, Chapra ad Caale, 5 th Ed., Part Fve, Chapter 8 ad
More informationAsymptotic Behavior of Solutions of Nonlinear Delay Differential Equations With Impulse
P a g e Vol Issue7Ver,oveber Global Joural of Scece Froer Research Asypoc Behavor of Soluos of olear Delay Dffereal Equaos Wh Ipulse Zhag xog GJSFR Classfcao - F FOR 3 Absrac Ths paper sudes he asypoc
More informationMachine Learning. Introduction to Regression. Lecture 3, September 19, Reading: Chap. 3, CB
ache Learg 0-70/5 70/5-78 78 all 006 Iroduco o Regresso Erc g Lecure 3 Sepember 9 006 Readg: Chap. 3 C Iferece wh he Jo Compue Codoals 0.4 0. P lu eadhead P lu eadhead P eadhead 0.7 0. 0.05 0.05 0.05 0.05
More informationTheory study about quarter-wave-stack dielectric mirrors
Theor tud about quarter-wave-tack delectrc rror Stratfed edu tratted reflected reflected Stratfed edu tratted cdet cdet T T Frt, coder a wave roagato a tratfed edu. A we kow, a arbtrarl olared lae wave
More information9.1 Introduction to the probit and logit models
EC3000 Ecoometrcs Lecture 9 Probt & Logt Aalss 9. Itroducto to the probt ad logt models 9. The logt model 9.3 The probt model Appedx 9. Itroducto to the probt ad logt models These models are used regressos
More informationSurvival Prediction Based on Compound Covariate under Cox Proportional Hazard Models
Ieraoal Bomerc Coferece 22/8/3, Kobe JAPAN Survval Predco Based o Compoud Covarae uder Co Proporoal Hazard Models PLoS ONE 7. do:.37/oural.poe.47627. hp://d.plos.org/.37/oural.poe.47627 Takesh Emura Graduae
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 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 informationA Second Kind Chebyshev Polynomial Approach for the Wave Equation Subject to an Integral Conservation Condition
SSN 76-7659 Eglad K Joural of forao ad Copug Scece Vol 7 No 3 pp 63-7 A Secod Kd Chebyshev olyoal Approach for he Wave Equao Subec o a egral Coservao Codo Soayeh Nea ad Yadollah rdokha Depare of aheacs
More informationMathematical Formulation
Mahemacal Formulao The purpose of a fe fferece equao s o appromae he paral ffereal equao (PE) whle maag he physcal meag. Eample PE: p c k FEs are usually formulae by Taylor Seres Epaso abou a po a eglecg
More informationGenerative classification models
CS 75 Mache Learg Lecture Geeratve classfcato models Mlos Hauskrecht mlos@cs.ptt.edu 539 Seott Square Data: D { d, d,.., d} d, Classfcato represets a dscrete class value Goal: lear f : X Y Bar classfcato
More informationChapter 2. Review of Hydrodynamics and Vector Analysis
her. Ree o Hdrodmcs d Vecor Alss. Tlor seres L L L L ' ' L L " " " M L L! " ' L " ' I s o he c e romed he Tlor seres. O he oher hd ' " L . osero o mss -dreco: L L IN ] OUT [mss l [mss l] mss ccmled h me
More informationBrownian Motion and Stochastic Calculus. Brownian Motion and Stochastic Calculus
Browa Moo Sochasc Calculus Xogzh Che Uversy of Hawa a Maoa earme of Mahemacs Seember, 8 Absrac Ths oe s abou oob decomoso he bascs of Suare egrable margales Coes oob-meyer ecomoso Suare Iegrable Margales
More informationCOMPARISON OF ESTIMATORS OF PARAMETERS FOR THE RAYLEIGH DISTRIBUTION
COMPARISON OF ESTIMATORS OF PARAMETERS FOR THE RAYLEIGH DISTRIBUTION Eldesoky E. Affy. Faculy of Eg. Shbee El kom Meoufa Uv. Key word : Raylegh dsrbuo, leas squares mehod, relave leas squares, leas absolue
More informationProbability Bracket Notation and Probability Modeling. Xing M. Wang Sherman Visual Lab, Sunnyvale, CA 94087, USA. Abstract
Probably Bracke Noao ad Probably Modelg Xg M. Wag Sherma Vsual Lab, Suyvale, CA 94087, USA Absrac Ispred by he Drac oao, a ew se of symbols, he Probably Bracke Noao (PBN) s proposed for probably modelg.
More informationSolution to Some Open Problems on E-super Vertex Magic Total Labeling of Graphs
Aalable a hp://paed/aa Appl Appl Mah ISS: 9-9466 Vol 0 Isse (Deceber 0) pp 04- Applcaos ad Appled Maheacs: A Ieraoal Joral (AAM) Solo o Soe Ope Probles o E-sper Verex Magc Toal Labelg o Graphs G Marh MS
More informationProbability and Statistics. What is probability? What is statistics?
robablt ad Statstcs What s robablt? What s statstcs? robablt ad Statstcs robablt Formall defed usg a set of aoms Seeks to determe the lkelhood that a gve evet or observato or measuremet wll or has haeed
More informationSome Different Perspectives on Linear Least Squares
Soe Dfferet Perspectves o Lear Least Squares A stadard proble statstcs s to easure a respose or depedet varable, y, at fed values of oe or ore depedet varables. Soetes there ests a deterstc odel y f (,,
More informationDomination in Controlled and Observed Distributed Parameter Systems
Iellge Cool ad Auoao 3 4 7-6 h://dxdoorg/436/ca346 Publshed Ole May 3 (h://wwwscrorg/joural/ca) Doao Coolled ad Observed Dsbued Paraeer yses L Aff M Joud E M Magr A El Ja Deare of Maheacs ad Couer cece
More informationIn 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 informationP a g e 3 6 of R e p o r t P B 4 / 0 9
P a g e 3 6 of R e p o r t P B 4 / 0 9 p r o t e c t h um a n h e a l t h a n d p r o p e r t y fr om t h e d a n g e rs i n h e r e n t i n m i n i n g o p e r a t i o n s s u c h a s a q u a r r y. J
More informationEMA5001 Lecture 3 Steady State & Nonsteady State Diffusion - Fick s 2 nd Law & Solutions
EMA5 Lecue 3 Seady Sae & Noseady Sae ffuso - Fck s d Law & Soluos EMA 5 Physcal Popees of Maeals Zhe heg (6) 3 Noseady Sae ff Fck s d Law Seady-Sae ffuso Seady Sae Seady Sae = Equlbum? No! Smlay: Sae fuco
More informationLecture 3 Topic 2: Distributions, hypothesis testing, and sample size determination
Lecure 3 Topc : Drbuo, hypohe eg, ad ample ze deermao The Sude - drbuo Coder a repeaed drawg of ample of ze from a ormal drbuo of mea. For each ample, compue,,, ad aoher ac,, where: The ac he devao of
More informationChapter 1 - Free Vibration of Multi-Degree-of-Freedom Systems - I
CEE49b Chaper - Free Vbrao of M-Degree-of-Freedo Syses - I Free Udaped Vbrao The basc ype of respose of -degree-of-freedo syses s free daped vbrao Aaogos o sge degree of freedo syses he aayss of free vbrao
More informationDensity estimation III. Linear regression.
Lecure 6 Mlos Hauskrec mlos@cs.p.eu 539 Seo Square Des esmao III. Lear regresso. Daa: Des esmao D { D D.. D} D a vecor of arbue values Obecve: r o esmae e uerlg rue probabl srbuo over varables X px usg
More informationOptimal Control and Hamiltonian System
Pure ad Appled Maheacs Joural 206; 5(3: 77-8 hp://www.scecepublshggroup.co//pa do: 0.648/.pa.2060503.3 ISSN: 2326-9790 (Pr; ISSN: 2326-982 (Ole Opal Corol ad Haloa Syse Esoh Shedrack Massawe Depare of
More informationInternet Appendix to: Idea Sharing and the Performance of Mutual Funds
Coes Iere Appedx o: Idea harg ad he Perforace of Muual Fuds Jule Cujea IA. Proof of Lea A....................................... IA. Proof of Lea A.3...................................... IA.3 Proof of
More information2/20/2013. Topics. Power Flow Part 1 Text: Power Transmission. Power Transmission. Power Transmission. Power Transmission
/0/0 Topcs Power Flow Part Text: 0-0. Power Trassso Revsted Power Flow Equatos Power Flow Proble Stateet ECEGR 45 Power Systes Power Trassso Power Trassso Recall that for a short trassso le, the power
More informationThe Mean Residual Lifetime of (n k + 1)-out-of-n Systems in Discrete Setting
Appled Mahemacs 4 5 466-477 Publshed Ole February 4 (hp//wwwscrporg/oural/am hp//dxdoorg/436/am45346 The Mea Resdual Lfeme of ( + -ou-of- Sysems Dscree Seg Maryam Torab Sahboom Deparme of Sascs Scece ad
More informationUpper Bound For Matrix Operators On Some Sequence Spaces
Suama Uer Bou formar Oeraors Uer Bou For Mar Oeraors O Some Sequece Saces Suama Dearme of Mahemacs Gaah Maa Uersy Yogyaara 558 INDONESIA Emal: suama@ugmac masomo@yahoocom Isar D alam aer aa susa masalah
More informationFresnel Equations cont.
Lecure 12 Chaper 4 Fresel quaos co. Toal eral refleco ad evaesce waves Opcal properes of meals Laer: Famlar aspecs of he eraco of lgh ad maer Fresel quaos r 2 Usg Sell s law, we ca re-wre: r s s r a a
More information-distributed random variables consisting of n samples each. Determine the asymptotic confidence intervals for
Assgme Sepha Brumme Ocober 8h, 003 9 h semeser, 70544 PREFACE I 004, I ed o sped wo semesers o a sudy abroad as a posgraduae exchage sude a he Uversy of Techology Sydey, Ausrala. Each opporuy o ehace my
More informationReliability Analysis. Basic Reliability Measures
elably /6/ elably Aaly Perae faul Œ elably decay Teporary faul Œ Ofe Seady ae characerzao Deg faul Œ elably growh durg eg & debuggg A pace hule Challeger Lauch, 986 Ocober 6, Bac elably Meaure elably:
More informationarxiv: v2 [math-ph] 3 Feb 2015
Nolocal ad global dyacs of cellular auoaa: A heorecal couer arhec for real couous as arxv:1312.6534v2 [ah-h] 3 Feb 2015 Vladr García-Morales 1 1 Isue for Advaced Sudy - Techsche Uversä Müche, Lchebergsr.
More informationCS 2750 Machine Learning Lecture 8. Linear regression. Supervised learning. a set of n examples
CS 75 Mache Learg Lecture 8 Lear regresso Mlos Hauskrecht los@cs.tt.eu 59 Seott Square Suervse learg Data: D { D D.. D} a set of eales D s a ut vector of sze s the esre outut gve b a teacher Obectve: lear
More informationSome probability inequalities for multivariate gamma and normal distributions. Abstract
-- Soe probably equales for ulvarae gaa ad oral dsrbuos Thoas oye Uversy of appled sceces Bge, Berlsrasse 9, D-554 Bge, Geray, e-al: hoas.roye@-ole.de Absrac The Gaussa correlao equaly for ulvarae zero-ea
More information1 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 informationAdvanced time-series analysis (University of Lund, Economic History Department)
Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng
More informationMultiple Choice Test. Chapter Adequacy of Models for Regression
Multple Choce Test Chapter 06.0 Adequac of Models for Regresso. For a lear regresso model to be cosdered adequate, the percetage of scaled resduals that eed to be the rage [-,] s greater tha or equal to
More informationCyclically Interval Total Colorings of Cycles and Middle Graphs of Cycles
Ope Joural of Dsree Mahemas 2017 7 200-217 hp://wwwsrporg/joural/ojdm ISSN Ole: 2161-7643 ISSN Pr: 2161-7635 Cylally Ierval Toal Colorgs of Cyles Mddle Graphs of Cyles Yogqag Zhao 1 Shju Su 2 1 Shool of
More information8. 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