2:k e-1x,- Y~ 12 /21(2,
|
|
- Susanna Gallagher
- 6 years ago
- Views:
Transcription
1 The Elastic et Apprach t the Traveling Salesman Prblem 349 C~~unicated by Andre~ Bart ~~~~~~~~~~~~~~~~~~~~ An Analysis f the Elastic et Apprach t the Traveling Salesman Prblem Richard Durbin* King's Cllege Rescnrcl: Centre, Cambridge CB 1ST, England Richard Szeliski Artificial Intelligence Center, SRI Internatinal, J\;fenl Park, CA 945 USA Alan Yuille Divisin f Applied Sciences, Harvard University, Cambridge, MA 138 USA This paper analyzes the elastic net apprach (Durbin and Willshaw 1987) t the traveling salesman prblem f finding the shrtest path thrugh a set f cities. The elastic net apprach intly minimizes the length f an arbitrary path in the plane and the distance between the path pints and the cities. The tradeff between these tw requirements is cntrlled by a scale parameter K. A glbal minimum is fund fr large K, and is then tracked t a small value. In this paper, we shw that (1) in the small K limit the elastic path passes arbitrarily clse t all the cities, but that nly ne path pint is attracted t each city, () in the large!< limit the net lies at the center f the set f cities, and (3) at a critical value f K the energy functin bifurcates. We als shw that this methd can be interpreted in terms f extremizing a prbability distributin cntrlled by K, The minimum at a given K crrespnds t the maximum a psteriri (MAP) Bayesian estimate f the tur under a natural statistical interpretatin. The analysis presented in this paper gives us a better understanding f the behavir f the elastic net, allws us t better chse the parameters fr the ptimizatin, and suggests hw t extend the underlying ideas t ther dmains. 1 Intrductin _ The traveling salesman prblem (Lawler et al. 1985) is a classical prblem in cmbinatrial ptimizatin. The task is t find the shrtest pssible tur thrugh a set f cities that passes thrugh each city exactly nce. This prblem is knwn t be P-cmplete, and it is generally believed "Current address: Department f Psychlgy, Stanfrd University, Stanfrd, CA 9435 USA. eural Cmputatin 1, (1989) 1989 Massachusetts Institute f Technlgy that the cmputatinal pwer needed t slve it grws expnentially with the number f cities. In this paper we analyze a recent parallel analg algrithm based n an elastic net. apprach (Durbin and Willshaw 1987) that generates gd slutins in much less time. A This apprach uses a fast heuristic methd with a strng gemetrical flavr that is based n the tea trader mdel f neural develpment (Willshaw and Vn der Malsburg 1979). It will wrk in a space f any dimensin, but fr simplicity we will assume the tw-dimensinal plane in this paper. Belw we briefly review the algrithm. Let {Xi}, i = 1 t, represent the psitins f the cities. The algrithm manipulates a path f pints in the plane, specified by [Y,}, =1 t M (AI larger than ), s that they eventually define a tur (that is, eventually each city Xi has sme path pint Y, cnverge t it). The path is updated each time step accrding t where ~ Y, = () L l1ji(xi - Y) + ;3!((Y Y-1 - Y ) i e-/x/- Y J 1 / 1( 71Ji =----- :k e-1x,- Y~ 1 /1(, () and (J are cnstants, and K is the scale parameter. Infrmally, the () term pulls the path tward the cities, s that fr each Xi there is at least ne Y, within distance apprximately K, The ;3 term pulls neighbring path pints tward each ther, and hence tries t make the path shrt. The update equatins are integrable, s that ~Y = -!(/DY fr an "energy" functin, E, given by E( {Y },!() = -(}1{ L lg L e- 1X,- Y J 1 / K + /JL {Y - y + } (1.1) 1 Fr fixed!\ the path will cnverge t a (pssibly lcal) minimum f E. At large values f K the energy functin is smthed and there is nly ne minimum. At small values f K, the energy functin cntains many lcal minima, all f which crrespnd t pssible turs f the cities (we prve this later in the paper), and the deepest minimum is the shrtest pssible tur. The algrithm. prceeds by starting at large K, and gradually reducing K, keeping t a lcal minimum f E (see Fig. 1). We wuld like this minimum that is tracked t remain the glbal minimum as ]\ becmes small. Unfrtunately, this can nt be guaranteed (see sectin 3). The elastic net apprach is similar t a number f previusly develped algrithms that use elastic matching (Burr 1981), energy-based matching (Terzpuls et al. 1987), r tpgraphic mapping (Khnen 1988) t slve visin, speech, and neural develpment prblems. Alternative parallel analg algrithms have als recently been prpsed fr slving the traveling salesman prblem (Hpfield and Tank 1985; Angenil et al. 1988). The methd f Angenil et al. is clsely related
2 35 Richard Durbin, Richard Szeliski, and Alan Yuille The Elastic et Apprach t the Traveling Salesman Prblem 351 t that discussed here, but is based n Khnen's self-rganizatin algrithm (Khnen 1988). It is faster, but fr large prblems it is marginally less accurate than the elastic methd. An imprtant cntributin f this paper is t analyze the behavir f the energy functin as the cnstant K changes and t describe the energy landscape. In particular, we prve results abut the behavir f the functin fr large and small K, cnfirming the assertins made abve abut hw' the algrithm wrks. First, hwever, we shw that (a) (b) (c) ~ ~ ~D (d) (e) (t) ~ {O Figure 1: The cnvergence f the netwrk as a functin f K, The white and black squares represent the data (1 cities) and the netwrk (5 path pints), respectively. The six figures (a-f) shw the cnfiguratin fund by the netwrk at values K =.61,.6,.1,.,.1,.4. The data set is centered n (.49,.46) and has secnd-rder mments (K;r;r,l(ry, K!J]) = (.75, -.3,.7). We use a =. and /3 =1.. The first bifurcatin, when the rigin becmes unstable, can be calculated t ccur at K =.66 (see Sectin 5), in agreement with the simulatin. The secnd break temperature is between K =.1 and K =. fr the simulatin when the line spreads int a lp. This crrespnds well t the pint at which the secnd eigenvalue becmes negative (K =.196). The crrespndence is nt exact because nnlinear terms becme significant after the first break. minimizing this cst functin can be interpreted in terms f maximizing a prbability distributin. The Prbabilistic Interpretatin _ The energy functin (1.1) can be related t a prbability distributin by expnentiatin. This is analgus t use f the Gibbs distributin in statistical mechanics. L({Y } }~) - 1 -E/K.r: - (7r)K Af e 1 A1 1 At II-{L ~e-ix'-yi/k}iie-nl;{{y-y+d i=l AI =l 7rK =l Observe that minimizing E with respect t {Y } crrespnds t maximizing L with respect t {Y }. We can interpret L in terms f Bayes' therem, which states that P(YIX) = P(XIY)P(Y) (.1) P(X) where P(YIX) is the prbability f a tur (Y) given a set f cities (X). Our algrithm maximizes P(YIX) ver all pssible turs (Y), s the value f P(X) is irrelevant. The distributin At P(Y) = IIp-11/K {Y - y./+1 } (.) )=1 is the a priri prbability f a given tur. This distributin is a crrelated gaussian that assigns greater prir prbability t shrter turs. The distributin 1 At 1 P(XIY) =II-{L ~e-ix'-yji/j{} (.3) ;=1 l~f )=1 7r1\ is the prbability f the cities being at (X) given that the tur pints are at (Y). P(XIY) is the prduct f IV independent prbability distributins P(XiIY) = ~ f= ~(,-IX'-YJI/1\ M )=1 7r1\ This equatin is equivalent t assuming that the measured psitin f city Xi was actually derived frm ne f the tur pints in {Y./} with a tw-dimensinal gaussian errr f variance 1(, but withut knwing which tur pint Xi crrespnds t. Thus equatin.1 shws that the elastic net algrithm is cmputing the mst prbable tur (finding the
3 35 Richard Durbin, Richard Szeliski, and Alan Yuille The Elastic et Apprach t the Traveling Salesman Prblem 353 Bayesian MAP estimate) given a prir mdel (.) that favrs shrt turs and a sensr mdel (.3) with tw-dimensinal psitin uncertainty. Our methd thus has an bvius extensin t surface interplatin and threedimensinal surface mdeling where the crrespndence between surface pints and measured data pints is unknwn. (a) 3 ~acking a ~ini~u~~~~~~~~~~~~~~~~~~~_ The algrithm devised by Durbin and Willshaw minimizes E at large!\ and then tracks the minimum energy slutin dwn t small K, At a lcal minimum (r any extremum), we have (J}/"./J =a 1 (e) (f) where II is 1 r and r? and ~ are the.1' and y cmpnents f the psitin vectr Y i As we fllw the extrema as K changes we get the equatin d () d!\ f)~/j =a (g) (i) which becmes f)e f)e drr./i D1\DYl' + Dy~fl in)" di~ = T btain the traectry we must slve this equatin fr d}~r [dl«, When we are at a true minimum, the Hessian D E DrT D1T is a psitive definite matrix and can be inverted, enabling us t cmpute rlrt [ili«, Bifurcatins ccur when the Hessian has zer eigenvalues. In this case dr~'l rl!\ is underdetermined and there are several pssible slutins. Cmputer simulatins and ur calculatins in the large!\ limit (see belw) shw that such a bifurcatin ccurs as the tur initially spreads ut frm a pint. After this, ur simulatins suggest that the minimum tracks smthly with /<...r. Other minima f the energy functin als appear as K is reduced. Fr the cnfiguratin shwn in Figure 1, the number f minima increases rapidly frm 1 at!{ =.1 t 3 at!{ =.1, 9 at!\ =.8 (shwn in Fig. ), and very many at!{ <.5. The minimum fund by tracking K frm large t small values is 11t necessarily the ptimal (glbal) minimum (Fig. ). evertheless, empirically the minima fund are within a few percent f ptimal (Durbin and Willshaw 1987). One pssible imprvement wuld be t pick up and track nearby minima by lcal randm perturbatin as in simulated annealing (Kirkpatrick et al. 1983). Figure : Pssible minima f the energy functin fr the data in Figure 1. T() investigate the energy minima we started 1 simulatins at /\r =.8 with randm initial cnfiguratins, hence 'withut using the slutins fr larger 1\ as initial cnditins. In cases that lead t a sensible tur n subsequent slw reductin f K, the lines ining the white bxes (data pints) shw the tur fund by the netwrk. We fund nine distinct grups f minima with the fllwing frequencies: (a) 6, (b) 183 (tur length 3.356), (c) 14 (tur length 3.88), (d) 13 (tur length 3.35), (e) 95, (f) 7, (g) 48, (h) 36 (tur length 3.4), and (i) 36. We have prbably fund all the mar frms f minima, since the least frequent happened 36 times. te that the mst frequent pattern (a) is neither the ne btained frm tracking /( (b) nr the ptimal ne (c). 4 The S~all K Li~it~~~~~~~~~~~~~~~~ We nw cnsider the behavir f the extrema and the Hessian at these extrema as 1\ -+ O. We will shw that the nly stable extrema ccur
4 354 Richard Durbin, Richard Szeliski, and Alan Yuille The Elastic et Apprach t the Traveling Salesman Prblem 355 when each {X;} has at least ne [Y} arbitrarily clse t it. Thus, the extrema all crrespnd t pssible turs. Fr any given i let B(!{) = min 1)[; - y~ I.J Then L e-ix/-yji//(! < J.~fe-1J(1\)/1\ and -O!\" lg ""'" e-ix'-yji/1\ > /"1 ~f B (!\")n G - -n \ g ~ +- K Thus, fr the energy t be bunded we must have. {B (!( ) - 1\"1 gai } C 11m = 1\ ----+!{ where (1 is a cnstant and hence B(I\") = ()(I\r 1 / ). Thus, in the limit as 1\" ~, cnfiguratins with unmatched Xs will have arbitrarily high energy, and s will nt be fund by the algrithm. This means that the minima will all crrespnd t pssible turs. Althugh all the Xs are matched there is n requirement that all the Ys are matched. Indeed, with crrect chice f parameters nand /1 it can be shwn that there will be nly ne tur pint at each city. The remaining tur pints space themselves evenly in the intercity intervals. A sufficient. requirement n the parameters fr this t happen is that J (\ <.4 where.4 is the shrtest distance frm a tur pint being attracted t sme city t any tur pint nt being attracted t that city. T derive this cnditin n the parameters cnsider a single city situated at the rigin. Define 1i'.J by Assume the Y, are at equilibrium. We wish t cnsider the stability f an equilibrium in which there are tw Ii'./ that stay significantly nnzer as!\" -----" (t have a single tur pint cnverge t each city we require instability). We can chse!\" sufficiently small that there is n significant interactin between these tw tur pints and any ther cities. Cnsider perturbatins Y = Y + Ei: We want t find eigenvectrs such that )..C = -/( DY' -nw~y + /11«Y~+l + Y~-l - Y~) ""'" DW..r -O{71'C + Y, G(-D. Ck)} + /11\O(f) + O(f ) k v, -a{1l'c + LY~(WWkY! - likw Y). cd + /J/{O(r) + ()«() k 1\ (11)" ""'".r = - ~ {Ii C + Y, G ldk'(y k. Ck) - Y(Y. C)} + /11i O(f) + ()«( ) 1\ k where f = max (I C I). In the limit f small 1{ we can ignre the /1 term as well as the higher rder f term. Clearly then fr each e = fy, fr sme scalar f' Let II = _),,1(/, and cnsider the case where nly il'l, 1i' are significant. This leads t the eigenvalue equatin 11 [/< - (1 - l1'l)r?] - II 11']11'1 ) ( f] ) = ( 1lJ]U)~ 11J [/( - (1 - l)}] - 11 f The criterin fr instability is that at least ne A is psitive, and hence that the crrespnding 11 is negative. Fr large 1( bth eigenvalues II] and 11 are clearly psitive. Hence fr instability we require that 1( shuld be belw the value at which ne eigenvalue (and hence their prduct) ges negative. The prduct is = 1111' [(1< - 11J }.~)(1( - r 71'1 Y'"l) - 11'11 y ] Y] n'1 11'1< [1( - Therefre we require that li r < 11' ~ + 11Jl y; (71' ~ + 11y;)] Since H'l + tr = 1 we will be safe if min IY I > 1(. At equilibrium at small!\" we have Y, = (;31\r/ u' )A where A = Y Y - 1. When, as will be usual, Y 1 and Y are neighbrs, then IA I is ust the distance t the next path pint nt cnverging n the city, and a sufficient requirement is that this distance must be greater than n /;1 Since an average tur n 1\T cities has length (/)1/ a safe estimate fr.a min wuld be.(i\t/)1/ / AI. Alternatively ne culd chse (\ t be a decreasing functin f 1(, such as 1{') where p is a fixed expnent between zer and ne. 5 At Large 1\" _ Fr large K, the energy functin (1.1) has a minimum crrespnding t the net lying at the center f the cities. At a critical value f 1\" this mini
5 356 Richard Durbin, Richard Szeliski, and Alan Yuille The Elastic et Apprach t the Traveling Salesman Prblem 357 mum becmes unstable and the system bifurcates. The initial mvement f the net depends nly n the secnd-rder mments f the cities. T shw this, we first calculate the first and secnd derivatives f E with respect t Y. () (Yl' - Xr)e-IX,-YA.1/J( = T'L ( At - DYl' \. i=l L=l e, + ;J{}TIJ _ vrt! _ VII } (5.1) k 1 k+1 1 k-1 D E DYl'DYt () 61IV6kle-IX,-YAI/J< KL M i=l (L=l e + n (Y{ - XJ')(}[V _ Xnf-IX,-Yd/J(f~IX'-YI1/J( }r3 L \. i=l : (}TI,J _ X!.J)(Y;v _ xr~)d-ix'-yki/j( ~,k 1 k 1 L U~'l K3 ~ (",~1 D-IXz-YJ I/J() 1,=1 L...,)=1 t + /J6 1W {6kl - 6~'+1l - 6k-1l} (5.) By substituting Y, = int (5.1) we find that _ (} (_Xf)e-IX,1/K _ (} IJ DYt - K L (L M e-1x,i/ K) - - K M LX, ~ 1=1 )=1 1=1 The rigin is thus always an extremum, prvided that it is chsen at the center f the Xs (i.e., Lt:1 Xi = ). T shw that the center is a minimum fr very large K, we must calculate the eigenvalues f the Hessian. As!( decreases, this minimum becmes unstable and a bifurcatin ccurs. Knwing the value f!( at. which this ccurs will give us a useful starting value fr!( when we are running the elastic net algrithm. At the rigin the Hessian can be written as D E _O_{)IJlJ{). +, XeX~) DYtDY~v AI!( U 1(3 A1 ~ ~{)kl L Xf'..X-r + ;1{)IIV {{)kl - {)k+ll - {)k-ll} (5.3) I\. ~.Al i=l Fr large K, the eigenvalues f the Hessian are clearly all psitive. By inspectin f equatin, we see that thrughut the regin IY I «!( the Hessian is psitive definite and s the rigin is a unique minimum. Fr small K, the dminant terms (the secnd and third terms n the right-hand side f 5.) have negative trace, s there are sme negative eigenvalues. Thus, the rigin is a stable state fr large K but then becmes unstable as K decreases. T see hw this ccurs we must explicitly calculate the eigenvectrs. If we cmpute the eigenvalues f the Hessian (Durbin et al. 1989), we find that smallest eigenvalue is. = _ ~ + J in' ~ (5 4 A mm tcai!<4ai 81 s AI. ) where A is the principal eigenvalue f the city cvariance matrix. The center then becmes unstable and breaks at!( s.t. Amin = O. This can be calculated frm equatin 5.4. Since the eigenvectrs depend nly n the secnd-rder mments f the distributin f the cities the glbal minimum fr K ust belw the critical value will als depend chiefly n the secnd-rder mments, As K decreases further the higher rder mments will becme imprtant. These theretical results are cnfirmed by cmputer simulatins (see Fig. 1). The net stays at the rigin until the critical value f K and then frms a line alng the principal axis f the city distributin. ear the secnd critical value f!( (when the eigenvalue determined by substituting the minr eigenvalue f the city cvariance matrix int equatin 5.4 becmes negative) the line spreads int a lp. 6 Cnclusin _ In summary, we have btained several theretical results cncerning the elastic net methd. First, we have shwn hw the elastic net slutin can be interpreted as a maximum a psteriri estimate f an unknwn tur (circular curve), where sme pints alng the tur have been measured with gaussian uncertainty in psitin. Secnd, we have prved that fr small K every pint Xi is matched, and that each pint must be within (!(1/) f a tur pint. Third, we have fund a cnditin n the parameters, IJ under which each city becmes matched by nly ne tur pint. Furth, we have shwn that at large K, a single minimum exists fr the energy functin, with all f the tur pints lying at the center f gravity f the cities. Fifth, we have shwn hw t calculate the bifurcatin pints fr the elastic net as K is reduced. The first result is particularly interesting since it suggests that this apprach can be applied t ther interplatin, apprximatin, and matching prblems (such as surface interplatin in cmputer visin). The imprtant feature here is that we d nt need t prespecify which mdel pint matches a data pint, allwing "slippery" matching. The secnd result prves the "crrectness" f the elastic net methd, in that any final slutin must be a valid tur. The third, furth, and fifth results can be used in selecting parameter values, a starting cnfiguratin fr the net, and a starting value fr Ii. The elastic net methd that we have analyzed prvides a simple, -effective, and intuitively satisfying algrithm fr generating gd traveling salesman turs. We believe that similar cntinuatin-based algrithms can be applied t a wide range f ptimizatin and apprximatin prblems.
Bootstrap Method > # Purpose: understand how bootstrap method works > obs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(obs) >
Btstrap Methd > # Purpse: understand hw btstrap methd wrks > bs=c(11.96, 5.03, 67.40, 16.07, 31.50, 7.73, 11.10, 22.38) > n=length(bs) > mean(bs) [1] 21.64625 > # estimate f lambda > lambda = 1/mean(bs);
More informationLyapunov Stability Stability of Equilibrium Points
Lyapunv Stability Stability f Equilibrium Pints 1. Stability f Equilibrium Pints - Definitins In this sectin we cnsider n-th rder nnlinear time varying cntinuus time (C) systems f the frm x = f ( t, x),
More informationinitially lcated away frm the data set never win the cmpetitin, resulting in a nnptimal nal cdebk, [2] [3] [4] and [5]. Khnen's Self Organizing Featur
Cdewrd Distributin fr Frequency Sensitive Cmpetitive Learning with One Dimensinal Input Data Aristides S. Galanpuls and Stanley C. Ahalt Department f Electrical Engineering The Ohi State University Abstract
More informationA Scalable Recurrent Neural Network Framework for Model-free
A Scalable Recurrent Neural Netwrk Framewrk fr Mdel-free POMDPs April 3, 2007 Zhenzhen Liu, Itamar Elhanany Machine Intelligence Lab Department f Electrical and Cmputer Engineering The University f Tennessee
More informationRevision: August 19, E Main Suite D Pullman, WA (509) Voice and Fax
.7.4: Direct frequency dmain circuit analysis Revisin: August 9, 00 5 E Main Suite D Pullman, WA 9963 (509) 334 6306 ice and Fax Overview n chapter.7., we determined the steadystate respnse f electrical
More informationChapter 3: Cluster Analysis
Chapter 3: Cluster Analysis } 3.1 Basic Cncepts f Clustering 3.1.1 Cluster Analysis 3.1. Clustering Categries } 3. Partitining Methds 3..1 The principle 3.. K-Means Methd 3..3 K-Medids Methd 3..4 CLARA
More informationComputational modeling techniques
Cmputatinal mdeling techniques Lecture 2: Mdeling change. In Petre Department f IT, Åb Akademi http://users.ab.fi/ipetre/cmpmd/ Cntent f the lecture Basic paradigm f mdeling change Examples Linear dynamical
More informationthe results to larger systems due to prop'erties of the projection algorithm. First, the number of hidden nodes must
M.E. Aggune, M.J. Dambrg, M.A. El-Sharkawi, R.J. Marks II and L.E. Atlas, "Dynamic and static security assessment f pwer systems using artificial neural netwrks", Prceedings f the NSF Wrkshp n Applicatins
More informationAdmissibility Conditions and Asymptotic Behavior of Strongly Regular Graphs
Admissibility Cnditins and Asympttic Behavir f Strngly Regular Graphs VASCO MOÇO MANO Department f Mathematics University f Prt Oprt PORTUGAL vascmcman@gmailcm LUÍS ANTÓNIO DE ALMEIDA VIEIRA Department
More informationMODULE 1. e x + c. [You can t separate a demominator, but you can divide a single denominator into each numerator term] a + b a(a + b)+1 = a + b
. REVIEW OF SOME BASIC ALGEBRA MODULE () Slving Equatins Yu shuld be able t slve fr x: a + b = c a d + e x + c and get x = e(ba +) b(c a) d(ba +) c Cmmn mistakes and strategies:. a b + c a b + a c, but
More informationThe blessing of dimensionality for kernel methods
fr kernel methds Building classifiers in high dimensinal space Pierre Dupnt Pierre.Dupnt@ucluvain.be Classifiers define decisin surfaces in sme feature space where the data is either initially represented
More informationSequential Allocation with Minimal Switching
In Cmputing Science and Statistics 28 (1996), pp. 567 572 Sequential Allcatin with Minimal Switching Quentin F. Stut 1 Janis Hardwick 1 EECS Dept., University f Michigan Statistics Dept., Purdue University
More informationAdmin. MDP Search Trees. Optimal Quantities. Reinforcement Learning
Admin Reinfrcement Learning Cntent adapted frm Berkeley CS188 MDP Search Trees Each MDP state prjects an expectimax-like search tree Optimal Quantities The value (utility) f a state s: V*(s) = expected
More informationWe can see from the graph above that the intersection is, i.e., [ ).
MTH 111 Cllege Algebra Lecture Ntes July 2, 2014 Functin Arithmetic: With nt t much difficulty, we ntice that inputs f functins are numbers, and utputs f functins are numbers. S whatever we can d with
More informationFall 2013 Physics 172 Recitation 3 Momentum and Springs
Fall 03 Physics 7 Recitatin 3 Mmentum and Springs Purpse: The purpse f this recitatin is t give yu experience wrking with mmentum and the mmentum update frmula. Readings: Chapter.3-.5 Learning Objectives:.3.
More informationResampling Methods. Cross-validation, Bootstrapping. Marek Petrik 2/21/2017
Resampling Methds Crss-validatin, Btstrapping Marek Petrik 2/21/2017 Sme f the figures in this presentatin are taken frm An Intrductin t Statistical Learning, with applicatins in R (Springer, 2013) with
More informationDifferentiation Applications 1: Related Rates
Differentiatin Applicatins 1: Related Rates 151 Differentiatin Applicatins 1: Related Rates Mdel 1: Sliding Ladder 10 ladder y 10 ladder 10 ladder A 10 ft ladder is leaning against a wall when the bttm
More informationx 1 Outline IAML: Logistic Regression Decision Boundaries Example Data
Outline IAML: Lgistic Regressin Charles Suttn and Victr Lavrenk Schl f Infrmatics Semester Lgistic functin Lgistic regressin Learning lgistic regressin Optimizatin The pwer f nn-linear basis functins Least-squares
More informationmaking triangle (ie same reference angle) ). This is a standard form that will allow us all to have the X= y=
Intrductin t Vectrs I 21 Intrductin t Vectrs I 22 I. Determine the hrizntal and vertical cmpnents f the resultant vectr by cunting n the grid. X= y= J. Draw a mangle with hrizntal and vertical cmpnents
More informationCHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS
CHAPTER 4 DIAGNOSTICS FOR INFLUENTIAL OBSERVATIONS 1 Influential bservatins are bservatins whse presence in the data can have a distrting effect n the parameter estimates and pssibly the entire analysis,
More informationMATHEMATICS SYLLABUS SECONDARY 5th YEAR
Eurpean Schls Office f the Secretary-General Pedaggical Develpment Unit Ref. : 011-01-D-8-en- Orig. : EN MATHEMATICS SYLLABUS SECONDARY 5th YEAR 6 perid/week curse APPROVED BY THE JOINT TEACHING COMMITTEE
More informationENSC Discrete Time Systems. Project Outline. Semester
ENSC 49 - iscrete Time Systems Prject Outline Semester 006-1. Objectives The gal f the prject is t design a channel fading simulatr. Upn successful cmpletin f the prject, yu will reinfrce yur understanding
More informationKinetic Model Completeness
5.68J/10.652J Spring 2003 Lecture Ntes Tuesday April 15, 2003 Kinetic Mdel Cmpleteness We say a chemical kinetic mdel is cmplete fr a particular reactin cnditin when it cntains all the species and reactins
More information, which yields. where z1. and z2
The Gaussian r Nrmal PDF, Page 1 The Gaussian r Nrmal Prbability Density Functin Authr: Jhn M Cimbala, Penn State University Latest revisin: 11 September 13 The Gaussian r Nrmal Prbability Density Functin
More informationNUMBERS, MATHEMATICS AND EQUATIONS
AUSTRALIAN CURRICULUM PHYSICS GETTING STARTED WITH PHYSICS NUMBERS, MATHEMATICS AND EQUATIONS An integral part t the understanding f ur physical wrld is the use f mathematical mdels which can be used t
More information4th Indian Institute of Astrophysics - PennState Astrostatistics School July, 2013 Vainu Bappu Observatory, Kavalur. Correlation and Regression
4th Indian Institute f Astrphysics - PennState Astrstatistics Schl July, 2013 Vainu Bappu Observatry, Kavalur Crrelatin and Regressin Rahul Ry Indian Statistical Institute, Delhi. Crrelatin Cnsider a tw
More informationLHS Mathematics Department Honors Pre-Calculus Final Exam 2002 Answers
LHS Mathematics Department Hnrs Pre-alculus Final Eam nswers Part Shrt Prblems The table at the right gives the ppulatin f Massachusetts ver the past several decades Using an epnential mdel, predict the
More informationk-nearest Neighbor How to choose k Average of k points more reliable when: Large k: noise in attributes +o o noise in class labels
Mtivating Example Memry-Based Learning Instance-Based Learning K-earest eighbr Inductive Assumptin Similar inputs map t similar utputs If nt true => learning is impssible If true => learning reduces t
More informationLeast Squares Optimal Filtering with Multirate Observations
Prc. 36th Asilmar Cnf. n Signals, Systems, and Cmputers, Pacific Grve, CA, Nvember 2002 Least Squares Optimal Filtering with Multirate Observatins Charles W. herrien and Anthny H. Hawes Department f Electrical
More informationPublic Key Cryptography. Tim van der Horst & Kent Seamons
Public Key Cryptgraphy Tim van der Hrst & Kent Seamns Last Updated: Oct 5, 2017 Asymmetric Encryptin Why Public Key Crypt is Cl Has a linear slutin t the key distributin prblem Symmetric crypt has an expnential
More informationPattern Recognition 2014 Support Vector Machines
Pattern Recgnitin 2014 Supprt Vectr Machines Ad Feelders Universiteit Utrecht Ad Feelders ( Universiteit Utrecht ) Pattern Recgnitin 1 / 55 Overview 1 Separable Case 2 Kernel Functins 3 Allwing Errrs (Sft
More informationLim f (x) e. Find the largest possible domain and its discontinuity points. Why is it discontinuous at those points (if any)?
THESE ARE SAMPLE QUESTIONS FOR EACH OF THE STUDENT LEARNING OUTCOMES (SLO) SET FOR THIS COURSE. SLO 1: Understand and use the cncept f the limit f a functin i. Use prperties f limits and ther techniques,
More informationT Algorithmic methods for data mining. Slide set 6: dimensionality reduction
T-61.5060 Algrithmic methds fr data mining Slide set 6: dimensinality reductin reading assignment LRU bk: 11.1 11.3 PCA tutrial in mycurses (ptinal) ptinal: An Elementary Prf f a Therem f Jhnsn and Lindenstrauss,
More informationSUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical model for microarray data analysis
SUPPLEMENTARY MATERIAL GaGa: a simple and flexible hierarchical mdel fr micrarray data analysis David Rssell Department f Bistatistics M.D. Andersn Cancer Center, Hustn, TX 77030, USA rsselldavid@gmail.cm
More informationENGI 4430 Parametric Vector Functions Page 2-01
ENGI 4430 Parametric Vectr Functins Page -01. Parametric Vectr Functins (cntinued) Any nn-zer vectr r can be decmpsed int its magnitude r and its directin: r rrˆ, where r r 0 Tangent Vectr: dx dy dz dr
More informationSection 6-2: Simplex Method: Maximization with Problem Constraints of the Form ~
Sectin 6-2: Simplex Methd: Maximizatin with Prblem Cnstraints f the Frm ~ Nte: This methd was develped by Gerge B. Dantzig in 1947 while n assignment t the U.S. Department f the Air Frce. Definitin: Standard
More informationPure adaptive search for finite global optimization*
Mathematical Prgramming 69 (1995) 443-448 Pure adaptive search fr finite glbal ptimizatin* Z.B. Zabinskya.*, G.R. Wd b, M.A. Steel c, W.P. Baritmpa c a Industrial Engineering Prgram, FU-20. University
More informationLecture 17: Free Energy of Multi-phase Solutions at Equilibrium
Lecture 17: 11.07.05 Free Energy f Multi-phase Slutins at Equilibrium Tday: LAST TIME...2 FREE ENERGY DIAGRAMS OF MULTI-PHASE SOLUTIONS 1...3 The cmmn tangent cnstructin and the lever rule...3 Practical
More informationThis section is primarily focused on tools to aid us in finding roots/zeros/ -intercepts of polynomials. Essentially, our focus turns to solving.
Sectin 3.2: Many f yu WILL need t watch the crrespnding vides fr this sectin n MyOpenMath! This sectin is primarily fcused n tls t aid us in finding rts/zers/ -intercepts f plynmials. Essentially, ur fcus
More information3.4 Shrinkage Methods Prostate Cancer Data Example (Continued) Ridge Regression
3.3.4 Prstate Cancer Data Example (Cntinued) 3.4 Shrinkage Methds 61 Table 3.3 shws the cefficients frm a number f different selectin and shrinkage methds. They are best-subset selectin using an all-subsets
More informationInterference is when two (or more) sets of waves meet and combine to produce a new pattern.
Interference Interference is when tw (r mre) sets f waves meet and cmbine t prduce a new pattern. This pattern can vary depending n the riginal wave directin, wavelength, amplitude, etc. The tw mst extreme
More informationDetermining the Accuracy of Modal Parameter Estimation Methods
Determining the Accuracy f Mdal Parameter Estimatin Methds by Michael Lee Ph.D., P.E. & Mar Richardsn Ph.D. Structural Measurement Systems Milpitas, CA Abstract The mst cmmn type f mdal testing system
More informationOF SIMPLY SUPPORTED PLYWOOD PLATES UNDER COMBINED EDGEWISE BENDING AND COMPRESSION
U. S. FOREST SERVICE RESEARCH PAPER FPL 50 DECEMBER U. S. DEPARTMENT OF AGRICULTURE FOREST SERVICE FOREST PRODUCTS LABORATORY OF SIMPLY SUPPORTED PLYWOOD PLATES UNDER COMBINED EDGEWISE BENDING AND COMPRESSION
More informationSupport-Vector Machines
Supprt-Vectr Machines Intrductin Supprt vectr machine is a linear machine with sme very nice prperties. Haykin chapter 6. See Alpaydin chapter 13 fr similar cntent. Nte: Part f this lecture drew material
More informationChapter 3 Kinematics in Two Dimensions; Vectors
Chapter 3 Kinematics in Tw Dimensins; Vectrs Vectrs and Scalars Additin f Vectrs Graphical Methds (One and Tw- Dimensin) Multiplicatin f a Vectr b a Scalar Subtractin f Vectrs Graphical Methds Adding Vectrs
More informationDistributions, spatial statistics and a Bayesian perspective
Distributins, spatial statistics and a Bayesian perspective Dug Nychka Natinal Center fr Atmspheric Research Distributins and densities Cnditinal distributins and Bayes Thm Bivariate nrmal Spatial statistics
More informationPhysics 2010 Motion with Constant Acceleration Experiment 1
. Physics 00 Mtin with Cnstant Acceleratin Experiment In this lab, we will study the mtin f a glider as it accelerates dwnhill n a tilted air track. The glider is supprted ver the air track by a cushin
More informationFigure 1a. A planar mechanism.
ME 5 - Machine Design I Fall Semester 0 Name f Student Lab Sectin Number EXAM. OPEN BOOK AND CLOSED NOTES. Mnday, September rd, 0 Write n ne side nly f the paper prvided fr yur slutins. Where necessary,
More informationComputational modeling techniques
Cmputatinal mdeling techniques Lecture 4: Mdel checing fr ODE mdels In Petre Department f IT, Åb Aademi http://www.users.ab.fi/ipetre/cmpmd/ Cntent Stichimetric matrix Calculating the mass cnservatin relatins
More informationSPH3U1 Lesson 06 Kinematics
PROJECTILE MOTION LEARNING GOALS Students will: Describe the mtin f an bject thrwn at arbitrary angles thrugh the air. Describe the hrizntal and vertical mtins f a prjectile. Slve prjectile mtin prblems.
More informationCS 477/677 Analysis of Algorithms Fall 2007 Dr. George Bebis Course Project Due Date: 11/29/2007
CS 477/677 Analysis f Algrithms Fall 2007 Dr. Gerge Bebis Curse Prject Due Date: 11/29/2007 Part1: Cmparisn f Srting Algrithms (70% f the prject grade) The bjective f the first part f the assignment is
More informationSlide04 (supplemental) Haykin Chapter 4 (both 2nd and 3rd ed): Multi-Layer Perceptrons
Slide04 supplemental) Haykin Chapter 4 bth 2nd and 3rd ed): Multi-Layer Perceptrns CPSC 636-600 Instructr: Ynsuck Che Heuristic fr Making Backprp Perfrm Better 1. Sequential vs. batch update: fr large
More informationx x
Mdeling the Dynamics f Life: Calculus and Prbability fr Life Scientists Frederick R. Adler cfrederick R. Adler, Department f Mathematics and Department f Bilgy, University f Utah, Salt Lake City, Utah
More informationMath Foundations 20 Work Plan
Math Fundatins 20 Wrk Plan Units / Tpics 20.8 Demnstrate understanding f systems f linear inequalities in tw variables. Time Frame December 1-3 weeks 6-10 Majr Learning Indicatrs Identify situatins relevant
More informationLead/Lag Compensator Frequency Domain Properties and Design Methods
Lectures 6 and 7 Lead/Lag Cmpensatr Frequency Dmain Prperties and Design Methds Definitin Cnsider the cmpensatr (ie cntrller Fr, it is called a lag cmpensatr s K Fr s, it is called a lead cmpensatr Ntatin
More informationIntroduction: A Generalized approach for computing the trajectories associated with the Newtonian N Body Problem
A Generalized apprach fr cmputing the trajectries assciated with the Newtnian N Bdy Prblem AbuBar Mehmd, Syed Umer Abbas Shah and Ghulam Shabbir Faculty f Engineering Sciences, GIK Institute f Engineering
More informationThermodynamics and Equilibrium
Thermdynamics and Equilibrium Thermdynamics Thermdynamics is the study f the relatinship between heat and ther frms f energy in a chemical r physical prcess. We intrduced the thermdynamic prperty f enthalpy,
More informationPreparation work for A2 Mathematics [2018]
Preparatin wrk fr A Mathematics [018] The wrk studied in Y1 will frm the fundatins n which will build upn in Year 13. It will nly be reviewed during Year 13, it will nt be retaught. This is t allw time
More informationPreparation work for A2 Mathematics [2017]
Preparatin wrk fr A2 Mathematics [2017] The wrk studied in Y12 after the return frm study leave is frm the Cre 3 mdule f the A2 Mathematics curse. This wrk will nly be reviewed during Year 13, it will
More informationSIZE BIAS IN LINE TRANSECT SAMPLING: A FIELD TEST. Mark C. Otto Statistics Research Division, Bureau of the Census Washington, D.C , U.S.A.
SIZE BIAS IN LINE TRANSECT SAMPLING: A FIELD TEST Mark C. Ott Statistics Research Divisin, Bureau f the Census Washingtn, D.C. 20233, U.S.A. and Kenneth H. Pllck Department f Statistics, Nrth Carlina State
More informationCOASTAL ENGINEERING Chapter 2
CASTAL ENGINEERING Chapter 2 GENERALIZED WAVE DIFFRACTIN DIAGRAMS J. W. Jhnsn Assciate Prfessr f Mechanical Engineering University f Califrnia Berkeley, Califrnia INTRDUCTIN Wave diffractin is the phenmenn
More informationBuilding to Transformations on Coordinate Axis Grade 5: Geometry Graph points on the coordinate plane to solve real-world and mathematical problems.
Building t Transfrmatins n Crdinate Axis Grade 5: Gemetry Graph pints n the crdinate plane t slve real-wrld and mathematical prblems. 5.G.1. Use a pair f perpendicular number lines, called axes, t define
More informationI. Analytical Potential and Field of a Uniform Rod. V E d. The definition of electric potential difference is
Length L>>a,b,c Phys 232 Lab 4 Ch 17 Electric Ptential Difference Materials: whitebards & pens, cmputers with VPythn, pwer supply & cables, multimeter, crkbard, thumbtacks, individual prbes and jined prbes,
More informationLab 1 The Scientific Method
INTRODUCTION The fllwing labratry exercise is designed t give yu, the student, an pprtunity t explre unknwn systems, r universes, and hypthesize pssible rules which may gvern the behavir within them. Scientific
More informationMATHEMATICS Higher Grade - Paper I
Higher Mathematics - Practice Eaminatin B Please nte the frmat f this practice eaminatin is different frm the current frmat. The paper timings are different and calculatrs can be used thrughut. MATHEMATICS
More informationROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS
Particle Acceleratrs, 1986, Vl. 19, pp. 99-105 0031-2460/86/1904-0099/$15.00/0 1986 Grdn and Breach, Science Publishers, S.A. Printed in the United States f America ROUNDING ERRORS IN BEAM-TRACKING CALCULATIONS
More informationCambridge Assessment International Education Cambridge Ordinary Level. Published
Cambridge Assessment Internatinal Educatin Cambridge Ordinary Level ADDITIONAL MATHEMATICS 4037/1 Paper 1 Octber/Nvember 017 MARK SCHEME Maximum Mark: 80 Published This mark scheme is published as an aid
More informationVerification of Quality Parameters of a Solar Panel and Modification in Formulae of its Series Resistance
Verificatin f Quality Parameters f a Slar Panel and Mdificatin in Frmulae f its Series Resistance Sanika Gawhane Pune-411037-India Onkar Hule Pune-411037- India Chinmy Kulkarni Pune-411037-India Ojas Pandav
More informationCOMP 551 Applied Machine Learning Lecture 11: Support Vector Machines
COMP 551 Applied Machine Learning Lecture 11: Supprt Vectr Machines Instructr: (jpineau@cs.mcgill.ca) Class web page: www.cs.mcgill.ca/~jpineau/cmp551 Unless therwise nted, all material psted fr this curse
More informationWhat is Statistical Learning?
What is Statistical Learning? Sales 5 10 15 20 25 Sales 5 10 15 20 25 Sales 5 10 15 20 25 0 50 100 200 300 TV 0 10 20 30 40 50 Radi 0 20 40 60 80 100 Newspaper Shwn are Sales vs TV, Radi and Newspaper,
More informationThermodynamics Partial Outline of Topics
Thermdynamics Partial Outline f Tpics I. The secnd law f thermdynamics addresses the issue f spntaneity and invlves a functin called entrpy (S): If a prcess is spntaneus, then Suniverse > 0 (2 nd Law!)
More informationA Simple Set of Test Matrices for Eigenvalue Programs*
Simple Set f Test Matrices fr Eigenvalue Prgrams* By C. W. Gear** bstract. Sets f simple matrices f rder N are given, tgether with all f their eigenvalues and right eigenvectrs, and simple rules fr generating
More informationDepartment of Economics, University of California, Davis Ecn 200C Micro Theory Professor Giacomo Bonanno. Insurance Markets
Department f Ecnmics, University f alifrnia, Davis Ecn 200 Micr Thery Prfessr Giacm Bnann Insurance Markets nsider an individual wh has an initial wealth f. ith sme prbability p he faces a lss f x (0
More informationFIELD QUALITY IN ACCELERATOR MAGNETS
FIELD QUALITY IN ACCELERATOR MAGNETS S. Russenschuck CERN, 1211 Geneva 23, Switzerland Abstract The field quality in the supercnducting magnets is expressed in terms f the cefficients f the Furier series
More informationFebruary 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA
February 28, 2013 COMMENTS ON DIFFUSION, DIFFUSIVITY AND DERIVATION OF HYPERBOLIC EQUATIONS DESCRIBING THE DIFFUSION PHENOMENA Mental Experiment regarding 1D randm walk Cnsider a cntainer f gas in thermal
More informationCONSTRUCTING STATECHART DIAGRAMS
CONSTRUCTING STATECHART DIAGRAMS The fllwing checklist shws the necessary steps fr cnstructing the statechart diagrams f a class. Subsequently, we will explain the individual steps further. Checklist 4.6
More informationDataflow Analysis and Abstract Interpretation
Dataflw Analysis and Abstract Interpretatin Cmputer Science and Artificial Intelligence Labratry MIT Nvember 9, 2015 Recap Last time we develped frm first principles an algrithm t derive invariants. Key
More informationPre-Calculus Individual Test 2017 February Regional
The abbreviatin NOTA means Nne f the Abve answers and shuld be chsen if chices A, B, C and D are nt crrect. N calculatr is allwed n this test. Arcfunctins (such as y = Arcsin( ) ) have traditinal restricted
More informationCHAPTER 3 INEQUALITIES. Copyright -The Institute of Chartered Accountants of India
CHAPTER 3 INEQUALITIES Cpyright -The Institute f Chartered Accuntants f India INEQUALITIES LEARNING OBJECTIVES One f the widely used decisin making prblems, nwadays, is t decide n the ptimal mix f scarce
More informationBLAST / HIDDEN MARKOV MODELS
CS262 (Winter 2015) Lecture 5 (January 20) Scribe: Kat Gregry BLAST / HIDDEN MARKOV MODELS BLAST CONTINUED HEURISTIC LOCAL ALIGNMENT Use Cmmnly used t search vast bilgical databases (n the rder f terabases/tetrabases)
More informationYou need to be able to define the following terms and answer basic questions about them:
CS440/ECE448 Sectin Q Fall 2017 Midterm Review Yu need t be able t define the fllwing terms and answer basic questins abut them: Intr t AI, agents and envirnments Pssible definitins f AI, prs and cns f
More information1996 Engineering Systems Design and Analysis Conference, Montpellier, France, July 1-4, 1996, Vol. 7, pp
THE POWER AND LIMIT OF NEURAL NETWORKS T. Y. Lin Department f Mathematics and Cmputer Science San Jse State University San Jse, Califrnia 959-003 tylin@cs.ssu.edu and Bereley Initiative in Sft Cmputing*
More informationKinetics of Particles. Chapter 3
Kinetics f Particles Chapter 3 1 Kinetics f Particles It is the study f the relatins existing between the frces acting n bdy, the mass f the bdy, and the mtin f the bdy. It is the study f the relatin between
More informationFlipping Physics Lecture Notes: Simple Harmonic Motion Introduction via a Horizontal Mass-Spring System
Flipping Physics Lecture Ntes: Simple Harmnic Mtin Intrductin via a Hrizntal Mass-Spring System A Hrizntal Mass-Spring System is where a mass is attached t a spring, riented hrizntally, and then placed
More informationMaximum A Posteriori (MAP) CS 109 Lecture 22 May 16th, 2016
Maximum A Psteriri (MAP) CS 109 Lecture 22 May 16th, 2016 Previusly in CS109 Game f Estimatrs Maximum Likelihd Nn spiler: this didn t happen Side Plt argmax argmax f lg Mther f ptimizatins? Reviving an
More informationEcology 302 Lecture III. Exponential Growth (Gotelli, Chapter 1; Ricklefs, Chapter 11, pp )
Eclgy 302 Lecture III. Expnential Grwth (Gtelli, Chapter 1; Ricklefs, Chapter 11, pp. 222-227) Apcalypse nw. The Santa Ana Watershed Prject Authrity pulls n punches in prtraying its missin in apcalyptic
More information(2) Even if such a value of k was possible, the neutrons multiply
CHANGE OF REACTOR Nuclear Thery - Curse 227 POWER WTH REACTVTY CHANGE n this lessn, we will cnsider hw neutrn density, neutrn flux and reactr pwer change when the multiplicatin factr, k, r the reactivity,
More informationEquilibrium of Stress
Equilibrium f Stress Cnsider tw perpendicular planes passing thrugh a pint p. The stress cmpnents acting n these planes are as shwn in ig. 3.4.1a. These stresses are usuall shwn tgether acting n a small
More informationChapter 2 GAUSS LAW Recommended Problems:
Chapter GAUSS LAW Recmmended Prblems: 1,4,5,6,7,9,11,13,15,18,19,1,7,9,31,35,37,39,41,43,45,47,49,51,55,57,61,6,69. LCTRIC FLUX lectric flux is a measure f the number f electric filed lines penetrating
More informationChapter Summary. Mathematical Induction Strong Induction Recursive Definitions Structural Induction Recursive Algorithms
Chapter 5 1 Chapter Summary Mathematical Inductin Strng Inductin Recursive Definitins Structural Inductin Recursive Algrithms Sectin 5.1 3 Sectin Summary Mathematical Inductin Examples f Prf by Mathematical
More informationNUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION
NUROP Chinese Pinyin T Chinese Character Cnversin NUROP CONGRESS PAPER CHINESE PINYIN TO CHINESE CHARACTER CONVERSION CHIA LI SHI 1 AND LUA KIM TENG 2 Schl f Cmputing, Natinal University f Singapre 3 Science
More informationANALYTICAL SOLUTIONS TO THE PROBLEM OF EDDY CURRENT PROBES
ANALYTICAL SOLUTIONS TO THE PROBLEM OF EDDY CURRENT PROBES CONSISTING OF LONG PARALLEL CONDUCTORS B. de Halleux, O. Lesage, C. Mertes and A. Ptchelintsev Mechanical Engineering Department Cathlic University
More informationMidwest Big Data Summer School: Machine Learning I: Introduction. Kris De Brabanter
Midwest Big Data Summer Schl: Machine Learning I: Intrductin Kris De Brabanter kbrabant@iastate.edu Iwa State University Department f Statistics Department f Cmputer Science June 24, 2016 1/24 Outline
More informationSurface and Contact Stress
Surface and Cntact Stress The cncept f the frce is fundamental t mechanics and many imprtant prblems can be cast in terms f frces nly, fr example the prblems cnsidered in Chapter. Hwever, mre sphisticated
More informationKinematic transformation of mechanical behavior Neville Hogan
inematic transfrmatin f mechanical behavir Neville Hgan Generalized crdinates are fundamental If we assume that a linkage may accurately be described as a cllectin f linked rigid bdies, their generalized
More informationResampling Methods. Chapter 5. Chapter 5 1 / 52
Resampling Methds Chapter 5 Chapter 5 1 / 52 1 51 Validatin set apprach 2 52 Crss validatin 3 53 Btstrap Chapter 5 2 / 52 Abut Resampling An imprtant statistical tl Pretending the data as ppulatin and
More informationFloating Point Method for Solving Transportation. Problems with Additional Constraints
Internatinal Mathematical Frum, Vl. 6, 20, n. 40, 983-992 Flating Pint Methd fr Slving Transprtatin Prblems with Additinal Cnstraints P. Pandian and D. Anuradha Department f Mathematics, Schl f Advanced
More informationThe Kullback-Leibler Kernel as a Framework for Discriminant and Localized Representations for Visual Recognition
The Kullback-Leibler Kernel as a Framewrk fr Discriminant and Lcalized Representatins fr Visual Recgnitin Nun Vascncels Purdy H Pedr Mren ECE Department University f Califrnia, San Dieg HP Labs Cambridge
More informationCOMP 551 Applied Machine Learning Lecture 5: Generative models for linear classification
COMP 551 Applied Machine Learning Lecture 5: Generative mdels fr linear classificatin Instructr: Herke van Hf (herke.vanhf@mail.mcgill.ca) Slides mstly by: Jelle Pineau Class web page: www.cs.mcgill.ca/~hvanh2/cmp551
More information