Average SIR of the desired user ([7]) β=0.99. β= Average NSE of the desired user (BADD) β=0.95. µ= β=0.9

Size: px
Start display at page:

Download "Average SIR of the desired user ([7]) β=0.99. β= Average NSE of the desired user (BADD) β=0.95. µ= β=0.9"

Transcription

1 A Blnd Adaptve Decorrelatng Detector for CDMA Systems Sennur Ulukus Roy D. Yates Department of lectrcal and Computer ngneerng Wreless Informaton Networks Laboratory (WINLAB) Rutgers Unversty, PO Box 909, Pscataway NJ Abstract: Te decorrelatng detector s known to elmnate te mult access nterference gven tat te sgnature sequences of te users are lnearly ndependent, at te cost of enancng te Gaussan recever nose. In ts paper, we present and study te convergence of a blnd adaptve decorrelatng detector wc s based on te observaton of te avalable statstcs. Te algortm s teratve and dstrbuted. Only two parameters are needed to be known for te constructon of te recever lter of a user: te user's sgnature sequence and te varance of te wte Gaussan recever nose. Te computatonal complexty of te proposed algortm per teraton s lnear n te number of users. 1 Introducton Code Dvson Multple Access (CDMA) systems suer from te near-far eect, because of non-ortogonalty of te users' sgnature sequences. Multuser detecton [1] can be used to overcome te near-far problem. Multuser detectors explot te specal structure of te multple access nterference to effectvely demodulate te non-ortogonal sgnals of te users. It was sown n [2] tat te optmal multuser detector as a computatonal complexty wc ncreases exponentally wt te number of actve users. Several suboptmum detectors, ncludng te decorrelatng detector [3], ave been proposed to aceve a performance comparable to tat of optmum detector wle keepng te complexty low. Te decorrelatng detector [3] wc as lnear (n number of users) computatonal complexty was sown to elmnate te mult access nterference totally f te sgnature sequences of te users are lnearly ndependent at te cost of enancng te addtve wte Gaussan nose (AWGN). Te decorrelatng detector of [3] s centralzed and non-teratve. Te constructon of te decorrelatng detector lter for a certan user requres te knowledge of te sgnature sequences of all nterferng users as well as te sgnature sequence of te user of nterest. In addton to tat, te constructon requres nverson of te N N cross correlaton matrx, were N s te number of actve users. Blnd adaptve algortms are desrable to overcome te need for te knowledge about te parameters of te nterferng users and teratve algortms are needed to avod te matrx nverson wc may ave a Supported by NSF Grant NCR large computatonal complexty. A blnd adaptve algortm based on te mnmzaton of te output energy was gven n [4]. Ts algortm was sown to converge to te MMS multuser detector [5]. In [6] an adaptve multuser detector wc converges to te decorrelatng detector s proposed. Ts detector stll needs te sgnature sequences of all te users and transmsson of tranng sequences. In [7] blnd algortms based on sgnal subspace trackng are nvestgated and two algortms wc converge to te decorrelatng and MMS multuser detectors are proposed. Te blnd adaptve decorrelatng detector proposed n [7] needs te nformaton about te varance of te AWGN and te number of users bot of wc can be estmated agan usng subspace trackng tecnques. Te computatonal complexty of te algortm of [7] s O(GN) per teraton, were G s te processng gan. In ts paper we present a blnd adaptve multuser detector wc uses observables tat are readly avalable at te recever and wc converges to a decorrelatng detector. Te detector s constructed va a dstrbuted teratve algortm wc updates te recever lter coecents of a desred user by usng te prevous output of te lter under constructon. Snce te lter output s random due to te randomness of te mult access nterference and exstence of ambent Gaussan cannel nose, te algortm evolves stocastcally. We prove te convergence of te lter coecents to a decorrelatng detector n te mean squared error (MS) sense. We develop lower and upper bounds for te MS of te lter coecents at teraton n + 1 n terms of te MS of te lter coecents at teraton n, and nvestgate te condtons under wc te MS sequence converges to zero as number of teratons grows to nnty. Te computatonal complexty of te proposed algortm s O(G) per teraton. Te proposed algortm requres te knowledge of only two parameters for te constructon of te lter coecents for a desred user: te desred user's sgnature sequence and te varance of te AWGN. Te varance of te AWGN s a xed quantty (not tme varyng) and can be estmated easly, peraps before te nformaton transfer starts wen only te samples of AWGN can be observed wtout any nterference at te output of an arbtrary (nonzero) recever lter. In suc a case, a relable estmate of te varance of AWGN can be obtaned by tme-averagng te squares of te output of te recever lter.

2 2 Te Decorrelatng Detector Trougout ts paper, a syncronous CDMA system wt BPSK modulaton s assumed to smplfy te analyss. We wll use G dmensonal vector s to denote te pre-assgned unque sgnature sequence of user. Let us dene an N G dmensonal matrx S wt ts ( j)t element beng (s ) j, te jt component of s. Terefore, te rows of S (equvalently te columns of S > ) are te sgnature sequences of te users. For future use, we dene a subspace L n G dmensonal vector space to be te subspace spanned by te sgnature sequences of te users,.e., L = spanfs 1 ::: s N g = column space of S > (1) We consder te decorrelatng detector for te rst user wtoutlossofgeneralty and represent ts G dmensonal recever lter by c. Ten c sould satsfy te followng condton. Sc = e 1 (2) were e 1 s te rst unt vector n N dmensonal space,.e., e 1 = [1 0 0 ::: 0] >, and s a non-negatve real number. It can be easly sown tat te bt error rate performance of te decorrelatng detector s nsenstve to scalng of te lter coecents as long as te lter elmnates te mult access nterference totally. Te reason for ts s tat te scalar factor multples bot te receved power level of te desred user and te AWGN. We rst note tat (2) as more tan one soluton, because t as N equaltes and G unknowns, and typcally G N. Te unque decorrelatng detector for te rst user, ~c, sgven [3] as ~c = S > (SS > ) ;1 e 1 (3) Let us denote any soluton of (2) as c. Ten nsertng (2) nto (3) we obtan (assumng =1) ~c = S > (SS > ) ;1 Sc = P c (4) Note tat P = S > (SS > ) ;1 S s te projecton matrx wc projects any vector onto te column space of S >. Also note tat te column space of S > s te subspace spanned by s 1 ::: s N, te subspace wc was denoted as L. Terefore, altoug (2) as more tan one soluton, all of te solutons ave te same projecton onto L and ts projecton s equal to te unque decorrelatng detector of Lupas and Verdu [3]. Let be an N N dmensonal dagonal matrx wt receved power of user, p,astst dagonal element. Multplyng bot sdes of (2) rst wt, ten wt S > yelds S > Sc = p 1 s 1 (5) We observe tat altoug (5) as G equatons n G unknowns t does not ave a unque soluton for c, snce (G ; N) egenvalues of S > S are equal to zero. Te soluton spaces of (2) and (5) are related as stated n te followng Remark. Remark 1 If s 1 ::: s N are lnearly ndependent, ten all solutons of (2) and (5) concde. Let us dene A = S > S. At ts pont we coose = 1=p 1 and devse te followng determnstc gradent descent algortm: c(n +1)=c(n) ; (Ac(n) ; s 1 ) (6) Note tat te teratve algortm gven n (6) converges to te soluton of (5) and equvalently, by Remark 1, to te soluton of (2). Te fact tat te soluton of (2) s not unque s noted earler. Note tat, for any c, PAc= S > (SS > ) ;1 SS > Sc = S > Sc = Ac (7) Ts means tat for any c(n), Ac(n) les entrely n L. Usng ts result and (6) we obtan c(n +1)=Pc(n +1)+c(n) ; Pc(n) (8) By nducton (8) yelds c(n) =Pc(n)+c(0) ; Pc(0) (9) Terefore, f te teratve algortm (6) s started n te subspace L, mplyng c(0) = Pc(0), ten from (9) we wll ave: c(n) =Pc(n) for all n. In ts case c(n) wll always stay n L and wll converge to te scaled verson of te unque decorrelatng detector soluton of [3] c = 1 p 1 ~c as n goes to nnty. Note tat te algortm converges to te scaled verson of ~c, nstead of convergng to ~c because s not cosen to be 1. Te restrcton tat te teratons sould be started n L for algortm (6) to converge to a decorrelatng detector s farly mld. Selectons c(0) = 0, c(0) = s 1 or any lnear combnaton of te sgnature sequences mply c(0) 2L, satsfyng te convergence condton of (6). Te sgnature sequences of all of te users mustbeknown for te algortm gven n (6). Also, te algortm of (6) s an o-lne algortm wc must be run before te real nformaton transmssons of te users start. After runnng te algortm for some tme, te lter coecents would converge to a decorrelatng detector lter and ten te communcaton can be started. In ts paper, our am s to develop a blnd adaptve algortm wc would converge to a decorrelatng detector soluton n a stocastc sense by usng real random measurements wle te users are actve and transmttng bts. To ts end, we wll propose an algortm wc can be vewed as te stocastc verson of te determnstc algortm gven n (6). 3 A Blnd Adaptve Decorrelatng Detector (BADD) Te receved base band sgnal at te nput of te recever lters after cp-matced lterng followed by cp rate samplng s NX p r = p b s + n (10) =1 were b s te nformaton bt (1 equprobably) transmtted by user and n s a Gaussan random vector wt zero mean

3 and [nn > ]= 2 I. Note tat, rr > = NX =1 p s s > + 2 I = A + 2 I (11) Terefore, usng (11), te determnstc teraton of (6) can be wrtten n an exact form as c(n +1)=c(n) ; ; rr > ; 2 I c(n) ; s 1 (12) Note tat rr > ; 2 I s an unbased estmate for A matrx. In order to obtan a practcal algortm we replace te exact expresson for A n (12) wt ts unbased estmator bysmply removng te expectaton n (12). Tus we obtan, c(n +1)=c(n) ; ; rr > ; 2 I c(n) ; s 1 (13) Before analyzng te convergence of (13), we state t n terms of avalable observatons, and lst te parameters needed at eac teraton. Let y(n) be te output of te recever lter of te desred user at tme n. Ten, y(n) =r > (n)c(n), were r(n) s te sampled receved sgnal before te recever lters and c(n) s te current lter of te desred user. Tus, te mplementaton orented verson of te algortm (13) s c(n +1)= ; 1 ; 2 c(n) ; (y(n)r(n) ; s 1 ) (14) Snce r(n) and y(n) are readly avalable at te nput and output of te recever lter tat s under constructon, only two system parameters are needed to be known n order to run te algortm: te sgnature sequence of te desred user s 1, and te varance of te AWGN, 2. Te varance of AWGN s a xed quantty wc can be easly estmated before te communcaton starts as dscussed n Secton 1. 4 Convergence of te BADD In ts Secton we wll nvestgate te convergence of te blnd adaptve decorrelatng detector proposed n te prevous secton. Let us dene te zero mean random vector (n) as (n) = ; rr > ; A ; 2 I c(n) (15) Notng tat Ac = s 1,we can wrte te stocastc teratons (13) and (14) as c(n +1)=c(n) ; [A (c(n) ; c )+(n)] (16) Subtractng c from bot sdes of (16) and denng (n) = c(n) ; c yelds (n +1)=(n) ; (A(n)+(n)) (17) Note tat norm of (n) s a measure of te dstance of te current recever lter from c, te convergence pont. Squarng of bot sdes of (17) yelds k(n +1)k 2 = k(n)k 2 ; 2(n) > A(n)+2 2 (n) > A(n) ;2(n) > (n)+ 2 (n) > A 2 (n)+ 2 k(n)k 2 (18) By takng te condtonal expectaton of bot sdes, condtoned on (n)=, and observng tat [(n) j (n)=]=0, we obtan k(n+1)k 2 j (n)= = kk 2 ; 2 > A + 2 > A k(n)k 2 j (n) = (19) We wll be usng te results of te followng Lemmas to develop bounds for te rgt and sde of (19). Lemma 1 If P = 0 ten > A = 0 and > A 2 = 0. If P6= 0 ten tere exst 0 <k 0 k 1 < 1, suctat k 0 kk 2 > A k 1 kk 2 (20) k 2 0 kk 2 > A 2 k 2 1 kk 2 (21) Lemma 2 Tere exst 0 <c 0 c 1 < 1, suc tat 0 k(n)k 2 j (n) = c 0 + c 1 kk 2 (22) Te outlne of te proof of Lemma 1 s as follows. Snce A s a postve semdente matrx wt rank N, (G ; N) egenvalues are equal to zero and remanng N egenvalues are postve. Snce for any x, Ax 2 L by (7), egenvectors of A are eter completely n L wt postve egenvalues, or completely out of L (meanng tat ter projectons onto L are zero) wt zero egenvalues. In ts case Lemma 1 s a smple result of Rayleg quotent [8]. Te proof of Lemma 2 wc can be found n [9] s more lengty and s omtted ere. In wat follows, we wll denote te condtonng on (n) =, P = 0 and P 6= 0 by, P and P, respectvely. For example [k(n +1)k 2 j (n) = P = 0] wll be denoted by [k(n +1)k 2 j P]. If P=0, followng lower and upper bounds can be developed for te rgt and sde of (19) usng Lemmas 1 and 2, k(n +1)k 2 j P k(n +1)k 2 j P And smlarly f P6= 0, k(n+1)k 2 j P ; 1+ 2 c 1 kk 2 + c 0 2 kk 2 (23) ; 1 ; 2k 0 +(k c 1 ) 2 kk 2 +c 0 2 k(n+1)k 2 j P ; 1 ; 2k 1 + k0 2 2 kk 2 (24) An mportant observaton toward te convergence proof s tat te projecton of c(n) (equvalently te projecton of (n)) on L would be non-zero almost surely (a.s.) [10]. Ts means tat for any n te probabltyofteevent[p(n) =0] s zero. Smlar to te determnstc result n (9), t can be sown usng nducton on (17) tat (see also [11, qn. (16)]), (n) =P(n)+(0) ; P(0) a.s. (25) Tus, f (0) 2Lten (0) = P(0) and (25) mples (n) = P(n) a.s. An upper bound for [k(n +1)k 2 j ] can be

4 developed as k(n+1)k 2 j = k(n+1)k 2 j P ProbfP = 0g + k(n+1)k 2 j P ProbfP6= 0g = k(n+1)k 2 j P ; 1 ; 2k 0 +(k c 1 ) 2 kk 2 +c 0 2 (26) were we used ProbfP = 0g = 0 and ProbfP 6= 0g =1. By smlar arguments a lower bound can be found to be k(n +1)k 2 j ; 1 ; 2k 1 + k kk 2 (27) Takng te expectaton of bot sdes of te nequaltes n (26) and (27) wt respect to (n) and lettng b n = [k(n)k 2 ], we obtan te followng bounds for b n, te mean squared error (MS) of te lter coecents at teraton n from c, b n+1 ; 1 ; 2k 1 + k b n (28) b n+1 ; 1 ; 2k 0 +(k c 1 ) 2 b n + c 0 2 (29) Denng, 0 =1;2k 0 +(k 2 1+c 1 ) 2 and 1 =1;2k 1 +k 2 0 2, we can rewrte quatons (28) and (29) as 1 b n b n+1 0 b n + c 0 2 (30) We observe from (30) tat te nonnegatve sequence b n s sandwced between te two sequences generated accordng to b 0 n+1 = 0 b 0 n+c 0 2 and b 00 n+1 = 1 b 00 n. Tese two sequences converge to nte numbers f and only f s cosen suc tat j 0 j< 1andj 1 j< 1. Note tat bot 0 and 1 are equal to1at =0. We also note tat bot 0 and 1 are locally decreasng as ncreases, snce d 0 d =0 = ;2k 0 < 0 and d 1 d =0 = ;2k 1 < 0 (31) Ts means tat we can always coose small enoug so tat j 0 j < 1 and j 1 j < 1, n wc case te sequences b 0 n and b 00 n converge and te lmtng MS,.e., lm n!1 b n, as nte lower and upper bounds. From te sandwc teorem, 0 lm n!1 b n c ; 0 (32) Te upper bound can be evaluated as! 0as, lm!0 c 0 2 c 0 = lm 1 ; 0!0 2k 0 ; (k1 2 + c =0 (33) 1) Terefore, f te step sze () s cosen extremely small ten te MS of te lter coecents from te decorrelatng lter coecents goes to zero as te number of teratons grows to nnty. But note tat as! 0, te numbers 0 and 1 go to 1 n wc case te convergence rate goes to zero. Tus, we observe te trade o between te lmtng value of te MS and te convergence rate. If a large value s cosen as te step sze ten te convergence rate s faster but te lmtng MS s larger and f a small value s cosen as te step sze te lmtng MS s smaller but te convergence rate s slower too. Hence, a tme dependent step sze sequence wc takes large values at te begnnng and smaller values at te end may be preferable. An teraton ndex (n) dependent step sze sequence can be used to accompls ts. Replacng te xed step sze n (14) wt te tme varyng step sze sequence a n we obtan te new algortm to be: c(n +1)= ; 1 ; 2 a n c(n) ; an (y(n)r(n) ; s 1 ) (34) Te convergence of (34) s stated n te followng Lemma. Te arguments of te proof wc can be found n [9] follow tose made n [12]. Lemma 3 If a n satses 1X n a n = 1 and 1X n a 2 n < 1 ten te stocastc teraton gven n (34) converges to c n te MS sense,.e., lm n!0 [kc(n) ; c k 2 ]=0. Note tat a n = a=(n + n 0 ), for any a>0 and n 0 > 0 satsfy te condtons of Lemma 3. 5 Smulaton Results In te smulatons we consder a system wt N = 6 users usng randomly generated sgnature sequences wt a processng gan of G =31. Powers of te users are assgned so tat te nal sgnal to nterference rato (SIR) of all users wll be 20 db. Tus, te power of te t user s found by p = [; ;1 ] were ; s te cross correlaton matrx wc s equal to SS >. Te algortm s run for 100 tmes and averaged results are plotted n te gures. In Fgure 1, we plot te averaged normalzed squared error (NS) of te lter coecents of te desred (rst) user versus teraton ndex n, were NS at teraton n s dened as NS(n) =kc(n) ; c k 2 = kc k 2. Varous curves n Fgure 1 correspond to te blnd adaptve decorrelatng detector algortms wt xed step sze for derent step sze values and tat wt a tme dependent step sze sequence of te form of a n = 1=(n + n 0 ) wt n 0 = 5. We observe tat f te step sze s too large, ten te algortm does not converge see ncreasng NS curve for =0:1 nfgure 1. We also observe tat for large step sze values te convergence rate s fast but te lmtng NS s large too compare te curves correspondng to =0:01 and =0:001 n Fgure 1. Averaged (over 100 runs) SIR of te desred (rst) user s plotted n Fgure 2 for te same system. At teraton n, te SIR of te desred user s calculated as p 1 (s > 1 SIR(n) = P c(n))2 p j6=1 j(s > j c(n))2 + 2 (c > (n)c(n)) (35) We observe tat te convergence propertes of te SIRs to te target SIR (wc s 20 db n ts experment) s smlar

5 to te propertes of te convergence of te NS to zero. Te blnd adaptve decorrelatng detector of [7] s also mplemented for te same system and te SIR of te desred user s plotted as a functon of teraton ndex n Fgure 3 for derent values of te forgettng factor β=0.999 Average SIR of te desred user ([7]) β=0.99 β= Average NS of te desred user (BADD) µ=0.1 Average SIR(n) 10 1 β=0.95 µ= β=0.9 Average NS(n) 10 1 µ=0.001 teraton ndex (n) µ=0.01 Fgure 3: Averaged SIR of te desred user (algortm of [7]). a n =1/(n+n 0 ) 10 2 teraton ndex (n) Fgure 1: Averaged normalzed squared error (NS) of te desred user. Average SIR(n) Average SIR of te desred user (BADD) a n =1/(n+n 0 ) µ=0.001 µ=0.01 µ= µ=0.1 teraton ndex (n) Fgure 2: Averaged SIR of te desred user. 6 Concluson In ts paper we presented an teratve and dstrbuted adaptve decorrelatng detector algortm wc s based on te observaton of te avalable statstcs, and studed ts convergence. Te update equaton of te algortm needs te observaton of te cp sampled nput sgnal before te recever lter and te output of te lter wt te current lter coecents. For te mplementaton of te algortm to construct te decorrelatng recever lter of a user only two parameters are needed to be known: te user's sgnature sequence and te varance of te addtve wte Gaussan recever nose. We studed te convergence of te proposed algortm bot for a xed step sze sequence and for a tme varyng step sze sequence. For te rst case we developed te condtons of avng lower and upper bounds on te MS and sowed tat as te step sze goes to zero te algortm converges n te MS. For te second case we drectly proved te convergence n te MS. References [1] S. Verdu. Multuser detecton. Advances n Statstcal Sgnal Processng, 2:369{409, [2] S. Verdu. Mnmum probablty of error for asyncronous gaussan multple-access cannels. I Trans. on Inform. Te., 32:85{96, January [3] R. Lupas and S. Verdu. Lnear multuser detectors for syncronous code-dvson multple-access cannels. I Trans. on Inform. Te., 35(1):123{136, January [4] M. Hong, U. Madow, and S. Verdu. Blnd adaptve multuser detecton. I Trans. on Inform. Te., 41(4):944{ 960, July [5] U. Madow and M. L. Hong. MMS nterference suppresson for drect-sequence spread-spectrum CDMA. I Trans. on Comm., 42(12):3178{3188, December [6] D. S. Cen and S. Roy. An adaptve multuser recever for CDMA systems. I JSAC, 12(5):808{816, June [7] X. Wang and H. V. Poor. Blnd adaptve nterference suppresson for CDMA communcatons based on egenspace trackng. In CISS-97, pages 468{473, Marc [8] G. Strang. Lnear Algebra and Its Applcatons. Saunders College Publsng, rd edton. [9] S. Ulukus and R. D. Yates. Convergence of a blnd adaptve decorrelatng detector. To be submtted. [10] W.F. Stout. Almost Sure Convergence. Academc Press, [11] L. Gyor. Adaptve lnear procedures under general condtons. I Trans. on Inform. Te., IT-30(2):262{267, Marc [12] D. J. Sakrson. Stocastc approxmaton : A recursve metod for solvng regresson problems. Advances n Comm. Systems 2, pages 51{106, A. V. Balakrsnan, d.

Multivariate Ratio Estimator of the Population Total under Stratified Random Sampling

Multivariate Ratio Estimator of the Population Total under Stratified Random Sampling Open Journal of Statstcs, 0,, 300-304 ttp://dx.do.org/0.436/ojs.0.3036 Publsed Onlne July 0 (ttp://www.scrp.org/journal/ojs) Multvarate Rato Estmator of te Populaton Total under Stratfed Random Samplng

More information

Stanford University CS254: Computational Complexity Notes 7 Luca Trevisan January 29, Notes for Lecture 7

Stanford University CS254: Computational Complexity Notes 7 Luca Trevisan January 29, Notes for Lecture 7 Stanford Unversty CS54: Computatonal Complexty Notes 7 Luca Trevsan January 9, 014 Notes for Lecture 7 1 Approxmate Countng wt an N oracle We complete te proof of te followng result: Teorem 1 For every

More information

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix

Lectures - Week 4 Matrix norms, Conditioning, Vector Spaces, Linear Independence, Spanning sets and Basis, Null space and Range of a Matrix Lectures - Week 4 Matrx norms, Condtonng, Vector Spaces, Lnear Independence, Spannng sets and Bass, Null space and Range of a Matrx Matrx Norms Now we turn to assocatng a number to each matrx. We could

More information

The Finite Element Method: A Short Introduction

The Finite Element Method: A Short Introduction Te Fnte Element Metod: A Sort ntroducton Wat s FEM? Te Fnte Element Metod (FEM) ntroduced by engneers n late 50 s and 60 s s a numercal tecnque for solvng problems wc are descrbed by Ordnary Dfferental

More information

Difference Equations

Difference Equations Dfference Equatons c Jan Vrbk 1 Bascs Suppose a sequence of numbers, say a 0,a 1,a,a 3,... s defned by a certan general relatonshp between, say, three consecutve values of the sequence, e.g. a + +3a +1

More information

Not-for-Publication Appendix to Optimal Asymptotic Least Aquares Estimation in a Singular Set-up

Not-for-Publication Appendix to Optimal Asymptotic Least Aquares Estimation in a Singular Set-up Not-for-Publcaton Aendx to Otmal Asymtotc Least Aquares Estmaton n a Sngular Set-u Antono Dez de los Ros Bank of Canada dezbankofcanada.ca December 214 A Proof of Proostons A.1 Proof of Prooston 1 Ts roof

More information

Composite Hypotheses testing

Composite Hypotheses testing Composte ypotheses testng In many hypothess testng problems there are many possble dstrbutons that can occur under each of the hypotheses. The output of the source s a set of parameters (ponts n a parameter

More information

Singular Value Decomposition: Theory and Applications

Singular Value Decomposition: Theory and Applications Sngular Value Decomposton: Theory and Applcatons Danel Khashab Sprng 2015 Last Update: March 2, 2015 1 Introducton A = UDV where columns of U and V are orthonormal and matrx D s dagonal wth postve real

More information

Y = X + " E [X 0 "] = 0 K E ["" 0 ] = = 2 I N : 2. So, you get an estimated parameter vector ^ OLS = (X 0 X) 1 X 0 Y:

Y = X +  E [X 0 ] = 0 K E [ 0 ] = = 2 I N : 2. So, you get an estimated parameter vector ^ OLS = (X 0 X) 1 X 0 Y: 1 Ecent OLS 1. Consder te model Y = X + " E [X 0 "] = 0 K E ["" 0 ] = = 2 I N : Ts s OLS appyland! OLS s BLUE ere. 2. So, you get an estmated parameter vector ^ OLS = (X 0 X) 1 X 0 Y: 3. You know tat t

More information

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 16

STAT 309: MATHEMATICAL COMPUTATIONS I FALL 2018 LECTURE 16 STAT 39: MATHEMATICAL COMPUTATIONS I FALL 218 LECTURE 16 1 why teratve methods f we have a lnear system Ax = b where A s very, very large but s ether sparse or structured (eg, banded, Toepltz, banded plus

More information

Competitive Experimentation and Private Information

Competitive Experimentation and Private Information Compettve Expermentaton an Prvate Informaton Guseppe Moscarn an Francesco Squntan Omtte Analyss not Submtte for Publcaton Dervatons for te Gamma-Exponental Moel Dervaton of expecte azar rates. By Bayes

More information

Error Probability for M Signals

Error Probability for M Signals Chapter 3 rror Probablty for M Sgnals In ths chapter we dscuss the error probablty n decdng whch of M sgnals was transmtted over an arbtrary channel. We assume the sgnals are represented by a set of orthonormal

More information

TR/95 February Splines G. H. BEHFOROOZ* & N. PAPAMICHAEL

TR/95 February Splines G. H. BEHFOROOZ* & N. PAPAMICHAEL TR/9 February 980 End Condtons for Interpolatory Quntc Splnes by G. H. BEHFOROOZ* & N. PAPAMICHAEL *Present address: Dept of Matematcs Unversty of Tabrz Tabrz Iran. W9609 A B S T R A C T Accurate end condtons

More information

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 8 Luca Trevisan February 17, 2016

U.C. Berkeley CS294: Spectral Methods and Expanders Handout 8 Luca Trevisan February 17, 2016 U.C. Berkeley CS94: Spectral Methods and Expanders Handout 8 Luca Trevsan February 7, 06 Lecture 8: Spectral Algorthms Wrap-up In whch we talk about even more generalzatons of Cheeger s nequaltes, and

More information

Throughput Capacities and Optimal Resource Allocation in Multiaccess Fading Channels

Throughput Capacities and Optimal Resource Allocation in Multiaccess Fading Channels Trougput Capactes and Optmal esource Allocaton n ultaccess Fadng Cannels Hao Zou arc 7, 003 Unversty of Notre Dame Abstract oble wreless envronment would ntroduce specal penomena suc as multpat fadng wc

More information

The Karush-Kuhn-Tucker. Nuno Vasconcelos ECE Department, UCSD

The Karush-Kuhn-Tucker. Nuno Vasconcelos ECE Department, UCSD e Karus-Kun-ucker condtons and dualt Nuno Vasconcelos ECE Department, UCSD Optmzaton goal: nd mamum or mnmum o a uncton Denton: gven unctons, g, 1,...,k and, 1,...m dened on some doman Ω R n mn w, w Ω

More information

Linear Approximation with Regularization and Moving Least Squares

Linear Approximation with Regularization and Moving Least Squares Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...

More information

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography

CSci 6974 and ECSE 6966 Math. Tech. for Vision, Graphics and Robotics Lecture 21, April 17, 2006 Estimating A Plane Homography CSc 6974 and ECSE 6966 Math. Tech. for Vson, Graphcs and Robotcs Lecture 21, Aprl 17, 2006 Estmatng A Plane Homography Overvew We contnue wth a dscusson of the major ssues, usng estmaton of plane projectve

More information

2.3 Nilpotent endomorphisms

2.3 Nilpotent endomorphisms s a block dagonal matrx, wth A Mat dm U (C) In fact, we can assume that B = B 1 B k, wth B an ordered bass of U, and that A = [f U ] B, where f U : U U s the restrcton of f to U 40 23 Nlpotent endomorphsms

More information

On Pfaff s solution of the Pfaff problem

On Pfaff s solution of the Pfaff problem Zur Pfaff scen Lösung des Pfaff scen Probles Mat. Ann. 7 (880) 53-530. On Pfaff s soluton of te Pfaff proble By A. MAYER n Lepzg Translated by D. H. Delpenc Te way tat Pfaff adopted for te ntegraton of

More information

bounds, but Mao [{4] only dscussed te mean square (te case of p = ) almost sure exponental stablty. Due to te new tecnques developed n ts paper, te re

bounds, but Mao [{4] only dscussed te mean square (te case of p = ) almost sure exponental stablty. Due to te new tecnques developed n ts paper, te re Asymptotc Propertes of Neutral Stocastc Derental Delay Equatons Xuerong Mao Department of Statstcs Modellng Scence Unversty of Stratclyde Glasgow G XH, Scotl, U.K. Abstract: Ts paper dscusses asymptotc

More information

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011

Stanford University CS359G: Graph Partitioning and Expanders Handout 4 Luca Trevisan January 13, 2011 Stanford Unversty CS359G: Graph Parttonng and Expanders Handout 4 Luca Trevsan January 3, 0 Lecture 4 In whch we prove the dffcult drecton of Cheeger s nequalty. As n the past lectures, consder an undrected

More information

On a nonlinear compactness lemma in L p (0, T ; B).

On a nonlinear compactness lemma in L p (0, T ; B). On a nonlnear compactness lemma n L p (, T ; B). Emmanuel Matre Laboratore de Matématques et Applcatons Unversté de Haute-Alsace 4, rue des Frères Lumère 6893 Mulouse E.Matre@ua.fr 3t February 22 Abstract

More information

Lecture Notes on Linear Regression

Lecture Notes on Linear Regression Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume

More information

Tracking with Kalman Filter

Tracking with Kalman Filter Trackng wth Kalman Flter Scott T. Acton Vrgna Image and Vdeo Analyss (VIVA), Charles L. Brown Department of Electrcal and Computer Engneerng Department of Bomedcal Engneerng Unversty of Vrgna, Charlottesvlle,

More information

Formulas for the Determinant

Formulas for the Determinant page 224 224 CHAPTER 3 Determnants e t te t e 2t 38 A = e t 2te t e 2t e t te t 2e 2t 39 If 123 A = 345, 456 compute the matrx product A adj(a) What can you conclude about det(a)? For Problems 40 43, use

More information

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0

n α j x j = 0 j=1 has a nontrivial solution. Here A is the n k matrix whose jth column is the vector for all t j=0 MODULE 2 Topcs: Lnear ndependence, bass and dmenson We have seen that f n a set of vectors one vector s a lnear combnaton of the remanng vectors n the set then the span of the set s unchanged f that vector

More information

Multigrid Methods and Applications in CFD

Multigrid Methods and Applications in CFD Multgrd Metods and Applcatons n CFD Mcael Wurst 0 May 009 Contents Introducton Typcal desgn of CFD solvers 3 Basc metods and ter propertes for solvng lnear systems of equatons 4 Geometrc Multgrd 3 5 Algebrac

More information

Population element: 1 2 N. 1.1 Sampling with Replacement: Hansen-Hurwitz Estimator(HH)

Population element: 1 2 N. 1.1 Sampling with Replacement: Hansen-Hurwitz Estimator(HH) Chapter 1 Samplng wth Unequal Probabltes Notaton: Populaton element: 1 2 N varable of nterest Y : y1 y2 y N Let s be a sample of elements drawn by a gven samplng method. In other words, s s a subset of

More information

A SINR Improvement Algorithm for D2D Communication Underlaying Cellular Networks

A SINR Improvement Algorithm for D2D Communication Underlaying Cellular Networks Advanced Scence and Tecnology Letters Vol.3 (CST 06), pp.78-83 ttp://dx.do.org/0.457/astl.06.3.34 A SINR Improvement Algortm for DD Communcaton Underlayng Cellular Networks Ceng uan, Youua Fu,, Jn Wang

More information

Digital Modems. Lecture 2

Digital Modems. Lecture 2 Dgtal Modems Lecture Revew We have shown that both Bayes and eyman/pearson crtera are based on the Lkelhood Rato Test (LRT) Λ ( r ) < > η Λ r s called observaton transformaton or suffcent statstc The crtera

More information

Statistical pattern recognition

Statistical pattern recognition Statstcal pattern recognton Bayes theorem Problem: decdng f a patent has a partcular condton based on a partcular test However, the test s mperfect Someone wth the condton may go undetected (false negatve

More information

MMA and GCMMA two methods for nonlinear optimization

MMA and GCMMA two methods for nonlinear optimization MMA and GCMMA two methods for nonlnear optmzaton Krster Svanberg Optmzaton and Systems Theory, KTH, Stockholm, Sweden. krlle@math.kth.se Ths note descrbes the algorthms used n the author s 2007 mplementatons

More information

Lecture 4: Constant Time SVD Approximation

Lecture 4: Constant Time SVD Approximation Spectral Algorthms and Representatons eb. 17, Mar. 3 and 8, 005 Lecture 4: Constant Tme SVD Approxmaton Lecturer: Santosh Vempala Scrbe: Jangzhuo Chen Ths topc conssts of three lectures 0/17, 03/03, 03/08),

More information

Multigradient for Neural Networks for Equalizers 1

Multigradient for Neural Networks for Equalizers 1 Multgradent for Neural Netorks for Equalzers 1 Chulhee ee, Jnook Go and Heeyoung Km Department of Electrcal and Electronc Engneerng Yonse Unversty 134 Shnchon-Dong, Seodaemun-Ku, Seoul 1-749, Korea ABSTRACT

More information

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal

Inner Product. Euclidean Space. Orthonormal Basis. Orthogonal Inner Product Defnton 1 () A Eucldean space s a fnte-dmensonal vector space over the reals R, wth an nner product,. Defnton 2 (Inner Product) An nner product, on a real vector space X s a symmetrc, blnear,

More information

The Order Relation and Trace Inequalities for. Hermitian Operators

The Order Relation and Trace Inequalities for. Hermitian Operators Internatonal Mathematcal Forum, Vol 3, 08, no, 507-57 HIKARI Ltd, wwwm-hkarcom https://doorg/0988/mf088055 The Order Relaton and Trace Inequaltes for Hermtan Operators Y Huang School of Informaton Scence

More information

Rethinking MIMO for Wireless Networks: Linear Throughput Increases with Multiple Receive Antennas

Rethinking MIMO for Wireless Networks: Linear Throughput Increases with Multiple Receive Antennas Retnng MIMO for Wreless etwors: Lnear Trougput Increases wt Multple Receve Antennas ar Jndal Unversty of Mnnesota Unverstat Pompeu Fabra Jont wor wt Jeff Andrews & Steven Weber MIMO n Pont-to-Pont Cannels

More information

5 The Laplace Equation in a convex polygon

5 The Laplace Equation in a convex polygon 5 Te Laplace Equaton n a convex polygon Te most mportant ellptc PDEs are te Laplace, te modfed Helmoltz and te Helmoltz equatons. Te Laplace equaton s u xx + u yy =. (5.) Te real and magnary parts of an

More information

Structure and Drive Paul A. Jensen Copyright July 20, 2003

Structure and Drive Paul A. Jensen Copyright July 20, 2003 Structure and Drve Paul A. Jensen Copyrght July 20, 2003 A system s made up of several operatons wth flow passng between them. The structure of the system descrbes the flow paths from nputs to outputs.

More information

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)

Econ107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4) I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes

More information

APPENDIX A Some Linear Algebra

APPENDIX A Some Linear Algebra APPENDIX A Some Lnear Algebra The collecton of m, n matrces A.1 Matrces a 1,1,..., a 1,n A = a m,1,..., a m,n wth real elements a,j s denoted by R m,n. If n = 1 then A s called a column vector. Smlarly,

More information

1 Introducton Nonlnearty crtera of Boolean functons Cryptograpc transformatons sould be nonlnear to be secure aganst varous attacks. For example, te s

1 Introducton Nonlnearty crtera of Boolean functons Cryptograpc transformatons sould be nonlnear to be secure aganst varous attacks. For example, te s KUIS{94{000 Nonlnearty crtera of Boolean functons HIROSE Souc IKEDA Katsuo Tel +81 75 753 5387 Fax +81 75 751 048 E-mal frose, kedag@kus.kyoto-u.ac.jp July 14, 1994 1 Introducton Nonlnearty crtera of Boolean

More information

The Geometry of Logit and Probit

The Geometry of Logit and Probit The Geometry of Logt and Probt Ths short note s meant as a supplement to Chapters and 3 of Spatal Models of Parlamentary Votng and the notaton and reference to fgures n the text below s to those two chapters.

More information

Econ Statistical Properties of the OLS estimator. Sanjaya DeSilva

Econ Statistical Properties of the OLS estimator. Sanjaya DeSilva Econ 39 - Statstcal Propertes of the OLS estmator Sanjaya DeSlva September, 008 1 Overvew Recall that the true regresson model s Y = β 0 + β 1 X + u (1) Applyng the OLS method to a sample of data, we estmate

More information

Research Article Green s Theorem for Sign Data

Research Article Green s Theorem for Sign Data Internatonal Scholarly Research Network ISRN Appled Mathematcs Volume 2012, Artcle ID 539359, 10 pages do:10.5402/2012/539359 Research Artcle Green s Theorem for Sgn Data Lous M. Houston The Unversty of

More information

CIS526: Machine Learning Lecture 3 (Sept 16, 2003) Linear Regression. Preparation help: Xiaoying Huang. x 1 θ 1 output... θ M x M

CIS526: Machine Learning Lecture 3 (Sept 16, 2003) Linear Regression. Preparation help: Xiaoying Huang. x 1 θ 1 output... θ M x M CIS56: achne Learnng Lecture 3 (Sept 6, 003) Preparaton help: Xaoyng Huang Lnear Regresson Lnear regresson can be represented by a functonal form: f(; θ) = θ 0 0 +θ + + θ = θ = 0 ote: 0 s a dummy attrbute

More information

More metrics on cartesian products

More metrics on cartesian products More metrcs on cartesan products If (X, d ) are metrc spaces for 1 n, then n Secton II4 of the lecture notes we defned three metrcs on X whose underlyng topologes are the product topology The purpose of

More information

COMP4630: λ-calculus

COMP4630: λ-calculus COMP4630: λ-calculus 4. Standardsaton Mcael Norrs Mcael.Norrs@ncta.com.au Canberra Researc Lab., NICTA Semester 2, 2015 Last Tme Confluence Te property tat dvergent evaluatons can rejon one anoter Proof

More information

The Minimum Universal Cost Flow in an Infeasible Flow Network

The Minimum Universal Cost Flow in an Infeasible Flow Network Journal of Scences, Islamc Republc of Iran 17(2): 175-180 (2006) Unversty of Tehran, ISSN 1016-1104 http://jscencesutacr The Mnmum Unversal Cost Flow n an Infeasble Flow Network H Saleh Fathabad * M Bagheran

More information

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U)

ANSWERS. Problem 1. and the moment generating function (mgf) by. defined for any real t. Use this to show that E( U) var( U) Econ 413 Exam 13 H ANSWERS Settet er nndelt 9 deloppgaver, A,B,C, som alle anbefales å telle lkt for å gøre det ltt lettere å stå. Svar er gtt . Unfortunately, there s a prntng error n the hnt of

More information

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.

Chapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise. Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where y + = β + β e for =,..., y and are observable varables e s a random error How can an estmaton rule be constructed for the

More information

Chapter 7 Channel Capacity and Coding

Chapter 7 Channel Capacity and Coding Wreless Informaton Transmsson System Lab. Chapter 7 Channel Capacty and Codng Insttute of Communcatons Engneerng atonal Sun Yat-sen Unversty Contents 7. Channel models and channel capacty 7.. Channel models

More information

, rst we solve te PDE's L ad L ad n g g (x) = ; = ; ; ; n () (x) = () Ten, we nd te uncton (x), te lnearzng eedbac and coordnates transormaton are gve

, rst we solve te PDE's L ad L ad n g g (x) = ; = ; ; ; n () (x) = () Ten, we nd te uncton (x), te lnearzng eedbac and coordnates transormaton are gve Freedom n Coordnates Transormaton or Exact Lnearzaton and ts Applcaton to Transent Beavor Improvement Kenj Fujmoto and Tosaru Suge Dvson o Appled Systems Scence, Kyoto Unversty, Uj, Kyoto, Japan suge@robotuassyoto-uacjp

More information

P R. Lecture 4. Theory and Applications of Pattern Recognition. Dept. of Electrical and Computer Engineering /

P R. Lecture 4. Theory and Applications of Pattern Recognition. Dept. of Electrical and Computer Engineering / Theory and Applcatons of Pattern Recognton 003, Rob Polkar, Rowan Unversty, Glassboro, NJ Lecture 4 Bayes Classfcaton Rule Dept. of Electrcal and Computer Engneerng 0909.40.0 / 0909.504.04 Theory & Applcatons

More information

Adaptive Kernel Estimation of the Conditional Quantiles

Adaptive Kernel Estimation of the Conditional Quantiles Internatonal Journal of Statstcs and Probablty; Vol. 5, No. ; 206 ISSN 927-7032 E-ISSN 927-7040 Publsed by Canadan Center of Scence and Educaton Adaptve Kernel Estmaton of te Condtonal Quantles Rad B.

More information

2 STATISTICALLY OPTIMAL TRAINING DATA 2.1 A CRITERION OF OPTIMALITY We revew the crteron of statstcally optmal tranng data (Fukumzu et al., 1994). We

2 STATISTICALLY OPTIMAL TRAINING DATA 2.1 A CRITERION OF OPTIMALITY We revew the crteron of statstcally optmal tranng data (Fukumzu et al., 1994). We Advances n Neural Informaton Processng Systems 8 Actve Learnng n Multlayer Perceptrons Kenj Fukumzu Informaton and Communcaton R&D Center, Rcoh Co., Ltd. 3-2-3, Shn-yokohama, Yokohama, 222 Japan E-mal:

More information

5 The Rational Canonical Form

5 The Rational Canonical Form 5 The Ratonal Canoncal Form Here p s a monc rreducble factor of the mnmum polynomal m T and s not necessarly of degree one Let F p denote the feld constructed earler n the course, consstng of all matrces

More information

Feature Selection: Part 1

Feature Selection: Part 1 CSE 546: Machne Learnng Lecture 5 Feature Selecton: Part 1 Instructor: Sham Kakade 1 Regresson n the hgh dmensonal settng How do we learn when the number of features d s greater than the sample sze n?

More information

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X

3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number

More information

CSCE 790S Background Results

CSCE 790S Background Results CSCE 790S Background Results Stephen A. Fenner September 8, 011 Abstract These results are background to the course CSCE 790S/CSCE 790B, Quantum Computaton and Informaton (Sprng 007 and Fall 011). Each

More information

The Prncpal Component Transform The Prncpal Component Transform s also called Karhunen-Loeve Transform (KLT, Hotellng Transform, oregenvector Transfor

The Prncpal Component Transform The Prncpal Component Transform s also called Karhunen-Loeve Transform (KLT, Hotellng Transform, oregenvector Transfor Prncpal Component Transform Multvarate Random Sgnals A real tme sgnal x(t can be consdered as a random process and ts samples x m (m =0; ;N, 1 a random vector: The mean vector of X s X =[x0; ;x N,1] T

More information

338 A^VÇÚO 1n ò Lke n Mancn (211), we make te followng assumpton to control te beavour of small jumps. Assumpton 1.1 L s symmetrc α-stable, were α (,

338 A^VÇÚO 1n ò Lke n Mancn (211), we make te followng assumpton to control te beavour of small jumps. Assumpton 1.1 L s symmetrc α-stable, were α (, A^VÇÚO 1n ò 1oÏ 215c8 Cnese Journal of Appled Probablty and Statstcs Vol.31 No.4 Aug. 215 Te Speed of Convergence of te Tresold Verson of Bpower Varaton for Semmartngales Xao Xaoyong Yn Hongwe (Department

More information

Channel Carrying: A Novel Hando Scheme. for Mobile Cellular Networks. Purdue University, West Lafayette, IN 47907, U.S.A.

Channel Carrying: A Novel Hando Scheme. for Mobile Cellular Networks. Purdue University, West Lafayette, IN 47907, U.S.A. Cannel Carryng: A Novel Hando Sceme for oble Cellular Networks Juny L Ness B. Sro y Edwn K.P. Cong Scool of Electrcal and Computer Engneerng Purdue Unversty, West Lafayette, IN 47907, U.S.A. E-mal: funy,

More information

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b

A New Refinement of Jacobi Method for Solution of Linear System Equations AX=b Int J Contemp Math Scences, Vol 3, 28, no 17, 819-827 A New Refnement of Jacob Method for Soluton of Lnear System Equatons AX=b F Naem Dafchah Department of Mathematcs, Faculty of Scences Unversty of Gulan,

More information

= = = (a) Use the MATLAB command rref to solve the system. (b) Let A be the coefficient matrix and B be the right-hand side of the system.

= = = (a) Use the MATLAB command rref to solve the system. (b) Let A be the coefficient matrix and B be the right-hand side of the system. Chapter Matlab Exercses Chapter Matlab Exercses. Consder the lnear system of Example n Secton.. x x x y z y y z (a) Use the MATLAB command rref to solve the system. (b) Let A be the coeffcent matrx and

More information

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction

The Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also

More information

Chat eld, C. and A.J.Collins, Introduction to multivariate analysis. Chapman & Hall, 1980

Chat eld, C. and A.J.Collins, Introduction to multivariate analysis. Chapman & Hall, 1980 MT07: Multvarate Statstcal Methods Mke Tso: emal mke.tso@manchester.ac.uk Webpage for notes: http://www.maths.manchester.ac.uk/~mkt/new_teachng.htm. Introducton to multvarate data. Books Chat eld, C. and

More information

Affine transformations and convexity

Affine transformations and convexity Affne transformatons and convexty The purpose of ths document s to prove some basc propertes of affne transformatons nvolvng convex sets. Here are a few onlne references for background nformaton: http://math.ucr.edu/

More information

Signal space Review on vector space Linear independence Metric space and norm Inner product

Signal space Review on vector space Linear independence Metric space and norm Inner product Sgnal space.... Revew on vector space.... Lnear ndependence... 3.3 Metrc space and norm... 4.4 Inner product... 5.5 Orthonormal bass... 7.6 Waveform communcaton system... 9.7 Some examples... 6 Sgnal space

More information

FAST CONVERGENCE ADAPTIVE MMSE RECEIVER FOR ASYNCHRONOUS DS-CDMA SYSTEMS

FAST CONVERGENCE ADAPTIVE MMSE RECEIVER FOR ASYNCHRONOUS DS-CDMA SYSTEMS Électronque et transmsson de l nformaton FAST CONVERGENCE ADAPTIVE MMSE RECEIVER FOR ASYNCHRONOUS DS-CDMA SYSTEMS CĂLIN VLĂDEANU, CONSTANTIN PALEOLOGU 1 Key words: DS-CDMA, MMSE adaptve recever, Least

More information

1 GSW Iterative Techniques for y = Ax

1 GSW Iterative Techniques for y = Ax 1 for y = A I m gong to cheat here. here are a lot of teratve technques that can be used to solve the general case of a set of smultaneous equatons (wrtten n the matr form as y = A), but ths chapter sn

More information

Supplement: Proofs and Technical Details for The Solution Path of the Generalized Lasso

Supplement: Proofs and Technical Details for The Solution Path of the Generalized Lasso Supplement: Proofs and Techncal Detals for The Soluton Path of the Generalzed Lasso Ryan J. Tbshran Jonathan Taylor In ths document we gve supplementary detals to the paper The Soluton Path of the Generalzed

More information

CME 302: NUMERICAL LINEAR ALGEBRA FALL 2005/06 LECTURE 13

CME 302: NUMERICAL LINEAR ALGEBRA FALL 2005/06 LECTURE 13 CME 30: NUMERICAL LINEAR ALGEBRA FALL 005/06 LECTURE 13 GENE H GOLUB 1 Iteratve Methods Very large problems (naturally sparse, from applcatons): teratve methods Structured matrces (even sometmes dense,

More information

VQ widely used in coding speech, image, and video

VQ widely used in coding speech, image, and video at Scalar quantzers are specal cases of vector quantzers (VQ): they are constraned to look at one sample at a tme (memoryless) VQ does not have such constrant better RD perfomance expected Source codng

More information

Appendix B: Resampling Algorithms

Appendix B: Resampling Algorithms 407 Appendx B: Resamplng Algorthms A common problem of all partcle flters s the degeneracy of weghts, whch conssts of the unbounded ncrease of the varance of the mportance weghts ω [ ] of the partcles

More information

1 Convex Optimization

1 Convex Optimization Convex Optmzaton We wll consder convex optmzaton problems. Namely, mnmzaton problems where the objectve s convex (we assume no constrants for now). Such problems often arse n machne learnng. For example,

More information

Problem Set 4: Sketch of Solutions

Problem Set 4: Sketch of Solutions Problem Set 4: Sketc of Solutons Informaton Economcs (Ec 55) George Georgads Due n class or by e-mal to quel@bu.edu at :30, Monday, December 8 Problem. Screenng A monopolst can produce a good n dfferent

More information

Application of Nonbinary LDPC Codes for Communication over Fading Channels Using Higher Order Modulations

Application of Nonbinary LDPC Codes for Communication over Fading Channels Using Higher Order Modulations Applcaton of Nonbnary LDPC Codes for Communcaton over Fadng Channels Usng Hgher Order Modulatons Rong-Hu Peng and Rong-Rong Chen Department of Electrcal and Computer Engneerng Unversty of Utah Ths work

More information

Convexity preserving interpolation by splines of arbitrary degree

Convexity preserving interpolation by splines of arbitrary degree Computer Scence Journal of Moldova, vol.18, no.1(52), 2010 Convexty preservng nterpolaton by splnes of arbtrary degree Igor Verlan Abstract In the present paper an algorthm of C 2 nterpolaton of dscrete

More information

Feb 14: Spatial analysis of data fields

Feb 14: Spatial analysis of data fields Feb 4: Spatal analyss of data felds Mappng rregularly sampled data onto a regular grd Many analyss technques for geophyscal data requre the data be located at regular ntervals n space and/or tme. hs s

More information

Eigenvalues of Random Graphs

Eigenvalues of Random Graphs Spectral Graph Theory Lecture 2 Egenvalues of Random Graphs Danel A. Spelman November 4, 202 2. Introducton In ths lecture, we consder a random graph on n vertces n whch each edge s chosen to be n the

More information

Errors for Linear Systems

Errors for Linear Systems Errors for Lnear Systems When we solve a lnear system Ax b we often do not know A and b exactly, but have only approxmatons  and ˆb avalable. Then the best thng we can do s to solve ˆx ˆb exactly whch

More information

Lecture 12: Classification

Lecture 12: Classification Lecture : Classfcaton g Dscrmnant functons g The optmal Bayes classfer g Quadratc classfers g Eucldean and Mahalanobs metrcs g K Nearest Neghbor Classfers Intellgent Sensor Systems Rcardo Guterrez-Osuna

More information

IN synchronous Code Division Multiple Access (CDMA)

IN synchronous Code Division Multiple Access (CDMA) 1 Optmal Groupng Algorthm for a Group Decson Feedbac Detector n Synchronous CDMA Communcatons J. Luo, K. Pattpat, P. Wllett, G. Levchu Abstract he Group Decson Feedbac (GDF) detector s studed n ths paper.

More information

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS

8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS SECTION 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS 493 8.4 COMPLEX VECTOR SPACES AND INNER PRODUCTS All the vector spaces you have studed thus far n the text are real vector spaces because the scalars

More information

General viscosity iterative method for a sequence of quasi-nonexpansive mappings

General viscosity iterative method for a sequence of quasi-nonexpansive mappings Avalable onlne at www.tjnsa.com J. Nonlnear Sc. Appl. 9 (2016), 5672 5682 Research Artcle General vscosty teratve method for a sequence of quas-nonexpansve mappngs Cuje Zhang, Ynan Wang College of Scence,

More information

OPTIMUM BEAMFORMING USING TRANSMIT ANTENNA ARRAYS. feasible points with monotonically decreasing costs).

OPTIMUM BEAMFORMING USING TRANSMIT ANTENNA ARRAYS. feasible points with monotonically decreasing costs). OPTIMUM BEAMFORMING USING TRANSMIT ANTENNA ARRAYS Eugene Vsotsky Upamanyu Madhow Abstract - Transmt beamformng s a powerful means of ncreasng capacty n systems n whch the transmtter s eupped wth an antenna

More information

e i is a random error

e i is a random error Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where + β + β e for,..., and are observable varables e s a random error How can an estmaton rule be constructed for the unknown

More information

Quantum Mechanics I - Session 4

Quantum Mechanics I - Session 4 Quantum Mechancs I - Sesson 4 Aprl 3, 05 Contents Operators Change of Bass 4 3 Egenvectors and Egenvalues 5 3. Denton....................................... 5 3. Rotaton n D....................................

More information

Consider the following passband digital communication system model. c t. modulator. t r a n s m i t t e r. signal decoder.

Consider the following passband digital communication system model. c t. modulator. t r a n s m i t t e r. signal decoder. PASSBAND DIGITAL MODULATION TECHNIQUES Consder the followng passband dgtal communcaton system model. cos( ω + φ ) c t message source m sgnal encoder s modulator s () t communcaton xt () channel t r a n

More information

Chapter 11: Simple Linear Regression and Correlation

Chapter 11: Simple Linear Regression and Correlation Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests

More information

The lower and upper bounds on Perron root of nonnegative irreducible matrices

The lower and upper bounds on Perron root of nonnegative irreducible matrices Journal of Computatonal Appled Mathematcs 217 (2008) 259 267 wwwelsevercom/locate/cam The lower upper bounds on Perron root of nonnegatve rreducble matrces Guang-Xn Huang a,, Feng Yn b,keguo a a College

More information

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems

Chapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons

More information

Outline. Multivariate Parametric Methods. Multivariate Data. Basic Multivariate Statistics. Steven J Zeil

Outline. Multivariate Parametric Methods. Multivariate Data. Basic Multivariate Statistics. Steven J Zeil Outlne Multvarate Parametrc Methods Steven J Zel Old Domnon Unv. Fall 2010 1 Multvarate Data 2 Multvarate ormal Dstrbuton 3 Multvarate Classfcaton Dscrmnants Tunng Complexty Dscrete Features 4 Multvarate

More information

Kernel Methods and SVMs Extension

Kernel Methods and SVMs Extension Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general

More information

EEE 241: Linear Systems

EEE 241: Linear Systems EEE : Lnear Systems Summary #: Backpropagaton BACKPROPAGATION The perceptron rule as well as the Wdrow Hoff learnng were desgned to tran sngle layer networks. They suffer from the same dsadvantage: they

More information

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin

LOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence

More information

Some modelling aspects for the Matlab implementation of MMA

Some modelling aspects for the Matlab implementation of MMA Some modellng aspects for the Matlab mplementaton of MMA Krster Svanberg krlle@math.kth.se Optmzaton and Systems Theory Department of Mathematcs KTH, SE 10044 Stockholm September 2004 1. Consdered optmzaton

More information

Some Comments on Accelerating Convergence of Iterative Sequences Using Direct Inversion of the Iterative Subspace (DIIS)

Some Comments on Accelerating Convergence of Iterative Sequences Using Direct Inversion of the Iterative Subspace (DIIS) Some Comments on Acceleratng Convergence of Iteratve Sequences Usng Drect Inverson of the Iteratve Subspace (DIIS) C. Davd Sherrll School of Chemstry and Bochemstry Georga Insttute of Technology May 1998

More information