Discussion Paper No Multivariate Time Series Model with Hierarchical Structure for Over-dispersed Discrete Outcomes

Size: px
Start display at page:

Download "Discussion Paper No Multivariate Time Series Model with Hierarchical Structure for Over-dispersed Discrete Outcomes"

Transcription

1 Dscusson Paper No. 113 Mulvarae Tme Seres Model wh Herarchcal Srucure for Over-dspersed Dscree Oucomes Nobuhko Teru and Masaaka Ban Augus, 213 January, 213 (Frs verson) TOHOKU MANAGEMENT & ACCOUNTING RESEARCH GROUP GRADUATE SCHOOL OF ECONOMICS AND MANAGEMENT TOHOKU UNIVERSITY KAWAUCHI, AOBA-KU, SENDAI JAPAN

2 Mulvarae Tme Seres Model wh Herarchcal Srucure for Over-dspersed Dscree Oucomes Nobuhko Teru * and Masaaka Ban ** Augus, 213 January, 213 *Graduae School of Economcs and Managemen, Tohoku Unversy, Senda , Japan **College of Economcs, Nhon Unversy, Chyoda-ku, Tokyo , Japan Correspondng Auhor: Nobuhko Teru, eru@econ.ohoku.ac.jp The auhors acknowledge useful commens from wo referees for revsng hs manuscrp. Teru also acknowledges he fnancal suppor of he Japanese Mnsry of Educaon Scenfc Research Grans (A)

3 Mulvarae Tme Seres Model wh Herarchcal Srucure for Over-dspersed Dscree Oucomes Absrac In hs paper, we propose a mulvarae me seres model for over-dspersed dscree daa o explore he marke srucure based on sales coun dynamcs. We frs dscuss he mcrosrucure o show ha over-dsperson s nheren n he modelng of marke srucure based on sales coun daa. The model s bul on he lkelhood funcon nduced by decomposng sales coun response varables accordng o producs compeveness and condonng on her sum of varables, and augmens hem o hgher levels by usng Posson-Mulnomal relaonshp n a herarchcal way, represened as a ree srucure for he marke defnon. Sae space prors are appled o he srucured lkelhood o develop dynamc generalzed lnear models for dscree oucomes. For over-dsperson problem, Gamma compound Posson varables for produc sales couns and Drchle compound mulnomal varables for her shares are conneced n a herarchcal fashon. Insead of he densy funcon of compound dsrbuons, we propose a daa augmenaon approach for more effcen poseror compuaons n erms of he generaed augmened varables parcularly for generang forecass and predcve densy. We presen he emprcal applcaon usng weekly produc sales me seres n a sore o compare he proposed models accommodang over-dsperson wh alernave no over-dspersed models by several model selecon crera, ncludng n-sample f, ou-of-sample forecasng errors, and nformaon creron. The emprcal resuls show ha he proposed modelng works well for he over-dspersed models based on compound Posson varables and hey provde mproved resuls han models wh no consderaon of over-dsperson. Key words: Compound Posson, Compound Mulnomal, Dscree Oucomes, Dynamc Generalzed Lnear Model, Herarchcal Marke Srucure, MCMC, Over-dsperson

4 1. Inroducon A full Bayesan analyss on he dynamcs of dscree responses such as couns has been faclaed by he Markov chan Mone Carlo (MCMC) mehods for more han 2 years. The scope s beyond he earler works wh maxmum lkelhood mehod by Harvey and Fernandes (1989) and Ord, Fernandes, and Harvey (1993). In parcular, Wes e al. (1985) and Cargnon e al. (1997) developed me-seres models for varables followng a mulnomal dsrbuon by nroducng dynamc lnear models usng he Bayesan approach. The former proposed a dynamc model wh a mulnomal dsrbuon and he laer deal wh several ses of mulnomal dsrbuons, boh of whch assumed he oal number of varables o be consan. In conras, he sochasc models for a dscree response exhb an neresng dsrbuonal propery: he reproducon of Posson varables and he condonal dsrbuon of hese varables on her sum follow a mulnomal dsrbuon. Teru e al. (21) used hs Posson mulnomal relaonshp n a dynamc generalzed lnear model o propose a mulvarae me-seres model wh a herarchcal srucure beween varables for specfyng marke srucure on he bass of a produc s sales dynamcs. We use he erm marke srucure o refer o how we classfy producs no several groups called submarkes or caegores, so ha he producs are compeve nsde a submarke bu no ousde. Ther model was a macro model for drec aggregae sales, whou consderaon of mcrosrucure. On he oher hand, he Posson varable has a lmed propery of havng dencal frs wo momens, and herefore he over-dsperson has been dscussed as mporan ssues n he leraure parcularly n economercs, as s fully dscussed n Wnkelmann (28). In hs paper, we exend he model of Teru e al. (21) such ha he dscree response varables have over-dspersons. We frs ncorporae he model of an ndvdual consumer s purchase and hen aggregae hose up o a produc sale, afer whch we fnd a mcrosrucure o generae over-dspersons for our applcaon. We prove ha over-dsperson s nheren 1

5 whenever consumers n he marke do no behave ndependenly, as s usually assumed n economcs and markeng. Then, we propose o buld a mulvarae me seres model wh herarchcal srucures by usng Gamma compound Posson varables for dscree responses havng over-dsperson. We frs develop he sascal modelng of compound dsrbuons o represen a marke srucure, and hen, propose a daa-augmenaon approach. Tha s, we rean he Posson Mulnomal dsrbuonal relaonshp mplyng he number of oal sales and produc s marke share, where we augmen Posson and Mulnomal parameers, whch are generaed by Gamma and Drchle dsrbuons, respecvely. Then, we oban he jon poseror densy by a full Bayesan MCMC procedure. We dscuss he mcrosrucure for he nheren over-dsperson n our problem n Secon 2. Secon 3 descrbes he properes of he compound Posson and compound mulnomal dsrbuons used as he buldng blocks for our model. The srucure of he model s explaned n Secon 4, and daa augmenaon approach o over-dsperson s proposed n Secon 5. Secon 6 descrbes model specfcaon n he dynamc generalzed lnear model and derves jon poseror densy. Secon 7 deals wh he esmaon and forecasng procedures. Secon 8 repors he emprcal applcaon. Concludng remarks are provded n Secon Mcrosrucure for Generang Over-dspersons Suppose ha here are H poenal consumers n he marke and he number of purchases of produc by consumer h a me follows a Posson dsrbuon wh parameer h as x h Posson, (1) h on he ground ha consumers wll make no purchase or small quany of buyng produc a a specfc me. Thus, he oal number of sales for produc, y, also follows Posson dsrbuon 2

6 H, (2) * h h h1 * h 1 H y x I hc x where I h C s an ndcaor funcon, akng value 1 when consumer h belongs o a poenal consumer se C,.e., when he/she s ready o buy, and oherwse, and H wh * H I h C h1 h beng he ndex for reorderng of consumers n he se C. In hs crcumsance, he over-dsperson phenomenon s derved by evaluang he mean and varance of y as because holds H H * * * h h, (3) E y E x * * h 1 h H y x * * * h x x h h ' * h hh' * * x * x * h h ' E y Var Var Cov, h * h * ' hh' Cov,, Cov x, x for any par of correlaed Posson varables. Tha s, equaon (4) mples ha here s over-dsperson whenever a leas one par of consumers does no behave ndependenly. Ths s no a srong assumpon, as jusfed by dscussons on he exsence of a reference group n socey and s decson makng, gong back o Hyman (1942), and s applcaon o socal psychology on he bass of he consumer behavor heory, such as Park and Lessg (1977) and Bearden and Ezel (1982). In fac, markeng models have been developed based on a commonaly across consumers when hey represen heerogeney n he random effec models. Ths s well explaned n he ex book Ross e al.(25). On he oher hand, Gamma compound Posson varables wh posve parameersa,, denoed as y~ Compound Posson, a, are suable for over-dsperson as hey conan (4) he frs wo momens 3

7 a ya Ey E y Var 1. (5) In he nex secon, we assume ha he number of produc sales, y, by aggregang over an ndvdual consumer s purchase, as shown n equaon (2), follows he compound Posson dsrbuon. 3. Gamma Compound Posson and Drchle Compound Mulnomal Dsrbuons A Gamma compound Posson varable s defned by he mxure of he Posson varable havng parameer n he Gamma dsrbuon wha,, and s densy funcon s evaluaed as a negave bnomal dsrbuon., a, p y a a a y 1 f y f a, d. y! 1 1 y a (6) Hoadley (1969) dscussed he reproducve propery of Gamma compound Posson varables and he condonal dsrbuon for a se of hese varables when he sum of varables s gven. Tha s, le y1, y2,..., y I be muually ndependen random varables havng a Gamma compound Posson dsrbuon wh he second parameer common across subjecs,.e., Then, he sum n y1 y2... yi follows y ~ Compound Posson a,. (7) n ~ Compound Posson, Furhermore, he condonal dsrbuon of y I a 1. (8) y1 y2 y I,,..., ' when n s gven s shown as y n~ Drchle Compound Mulnomal y n, a, (9) 4

8 where a a I, 1,..., and s densy s derved by n 1 n! a y a py n, a f y n, f ad. (1) n a y! a 4. Models for Defnng Marke Srucure The producs are more or less compeve n her marke. These producs are grouped accordng o he degree of compeveness no several segmens, and furher caegorzng hese groups o hgher levels o make subgroups leads o a herarchcal srucure for he marke defnon. A ree srucure s used o represen he herarchcal naure of compeve relaonshps among producs n a graphcal form, as shown n Fgure 1. Fgure 1(a) ndcaes he marke wh no specfc srucure and each produc sale s drecly accumulaed o marke sale. On he oher hand, Fgure 1(b) shows he suaon ha some groups of produc sales are respecvely accumulaed o sub-markes frs and hen sub marke sales consue marke sale. Fgure 1 (a), (b) Marke Srucures Basc Srucure Le us assume ha here are I producs n he marke and ha y s he number of sales for he produc a me ( 1,..., T), whch follows he Gamma compound Posson dsrbuon ndependenly wh a me-varyng parameer a, defned by (7) for 1,..., I when here s no compeve relaonshp wh each oher. Then, we oban he Gamma compound Posson dsrbuon wh a, for marke sales, defned as he aggregae of produc sales, n I y 1, under he assumpon of no specfc srucure among producs n he marke, as shown n Fgure 1(a). Tha s, we have margnal dsrbuons for 5

9 produc and marke sales by * y ~ Compound Posson a,, n Compound Posson a,, (11) where a * I a. Furhermore, afer 1 n s gven, he condonal dsrbuon of produc follows sales y y, 1,..., I where a a a y n~ Drchle Compound Mulnomal y n, a, (12),..., ' 1. The sequenal use of equaons (11) and (12) produces a jon I dsrbuon for marke and produc sales, *,,,, p n y a p n a p y n a. (13) We noe ha he condonng se a *,, a has equvalen nformaon wh a, f we ake a as a full-dmensonal vecor wh a nondegeneraed dsrbuon. In conras, Teru e al. (21) proposed a dynamc generalzed lnear model based on Posson varables whou over-dspersons. Ths represens a macromodel for aggregae sales drecly whou consderng he mcrosrucure. They used he reproducve propery of Posson varables and condonal mulnomal dsrbuon when he sum of varables s gven, and proposed a mulvarae me-seres model wh a herarchcal srucure based on he dscree oucomes. Tha s, we have margnal dsrbuons y Posson and n * Posson and condonal dsrbuon y n Mulnomal y n,, where * I I and j j, j 1,..., I 1 1. The lkelhood a me s defned by 6

10 * where he nduced parameers and and marke shares for each produc. Hgher Order Srucure * *,,, p n y p n p y n, (14), respecvely, represen he expeced oal sales Ths model s exended o a hgher order herarchcal srucure, as developed by Teru e al. (21). Nex, we fully explan he model specfcaons, ncludng he desgn marx of sae space prors, as hs model s appled o acual me-seres daa n he emprcal analyss. Exendng he Posson mulnomal relaonshp n he above-descrbed manner, we decompose he marke srucure no L submarkes [ k ], ' k M k such ha y L k1 y [ k ], where y y M represens N k dmensonal vecor of he producs ha are grouped n M, k 1,..., L. Gven aggregaed submarke sales k m [ k ] Mk y, he condonal [ k] [ k] dsrbuon y m, k 1,..., L, follows ndependen N k -dmensonal mulnomal [ k ] dsrbuon snce l for [ l] [ l] [ k] [ k] y s are orhogonal o each oher,.e., y m y m k by he defnon of a submarke.,..., ' denoe an L-dmensonal vecor of submarke sales. Then, [1] [ L] Nex, le m m m m n follows a mulnomal dsrbuon condonal on he sum of submarke sales,.e., marke sales n We noe ha L m k 1 [ k ]. In bref, we have a hree-layer herarchcal marke srucure model. y and n are ndependen, condonal on m. Then, he jon densy funcon of I produc sales (boom layer), L submarke sales (mddle layer), and marke sales (op layer) are decomposed no,, p n m y p n p m n p y m L [ k] [ k] [ k] p n p m n p y m. k 1 (15) 7

11 , * Then, we oban margnal and condonal daa dsrbuons: n Posson m n Mulnomal n, y m Mulnomal m,, where k L [ k ] [ k ] [ k ] [ k ] N k [ k] [ k] [ k] / j j1 N j L Nk [ j] [ k] j / 1 k1 1 j 1,..., L 1 [ k] [ k 1,...,, ], M k,, 1,..., 1, and [ k ] s he parameer of he Posson varable N k classfed o he submarke M k. The srucure of hs model s llusraed n Fgure 1(b). 5. Daa Augmenaon Approach o Over-dsperson Daa Augmenaon Exendng he model for accommodang over-dsperson, nsead of drec use of denses (6) and (1), we ake a daa augmenaon approach o keep he orgnal parameers,.e., mplyng expeced sales and marke shares for each produc n he modelng, and use he * ' generaed sample of augmened varable z lkelhood for he parameersa, :, ' n he MCMC process o defne he *, =Gamma, Drchle p z a a a. (16) By usng he relaon of poseror densy * *,,,,,, * * * *,, p a n y p a n y d d (17) p n p a d p y n p a d, we evaluae hese negrals by augmenng he gamma and Drchle parameers n erms of generang he s-h samples *( s ) *( ) *( ) ( ) from gamma p s s, s a and ( s) from Drchle p ( s ) ( s ) ( s) *( s) ( s) n, a n MCMC eraons. Then, condonal on z oban he Posson Mulnomal lkelhood funcon, ' ', we 8

12 *( s) ( s) pn p y n,, (18) and hs forms a buldng block o consue a herarchcal srucure. In case of wo-layer models and hree submarke (L=3), s exended as follows: for * ' [1] [2] [3] z,, ', ', ' ',,,,, ;,, * * * * pn p a, d p a n m y p a z n m y dz L p m n p a d p y m p a d [ k] [ k] [ k] [ k] [ k] [ k], m,, k 1 (19) where am a, k 1,..., L Mk [ k ] and, a a M. k We evaluae hese negrals by augmenng he gamma and Drchle parameers erms of generaed samples ( s ( ) Drchle ) s p n, am *( s ) *( ) *( ) ( ) from Gamma p s s, s a and ( ) ( ) [ k ] s [ k ] s p a ( s) *( s) '( s) [1]( s) [2]( s) [3]( s), ( s) and [ k]( s) z n from, respecvely. Then, condonal on z,, ', ', ' ', we oban he Posson Mulnomal lkelhood funcon: L *( s ) ( s ) [ k ]( s ) [ k, ]( s ), [ k ]( s ) p n p m n p y m. (2) k 1 6. Model Specfcaon and Jon Poseror Densy Dynamc Generalzed Lnear Model Usng he expecaon of produc sales leads o E y a. (21) Thus, we nerpre ha he expeced sale s decomposed no an ndvdual mean a and a common mean across producs. In urn, we model he mean funcon as a f x, 9

13 connecng wh markeng mx varables x and sochasc error. We specfy he srucure n more deal as log a a log x ', 1,..., I. (22) Tha s, he ndvdual mean has a produc-specfc mean a and a common me rend log, * and s assumed o be mosly explaned by markeng mx varables. We se a log a a * and denoe log, and hen condonal on a, we have he dynamc equaon a * * Ths consues he srucural equaon F v * = 1, 2,..., I, ' x '. (23) * *, where a ai, F s he marx defned by he srucure on 1,..., ' and n equaon (24), and v s he error vecor comprsng of. As for he dynamc sae vecor used n he * applcaon, we specfy he second-order local common rend for and he frs-order local rend model for he response parameer,.e., w ; w 1w3, 1,... I, * * (24) and hs specfes he sysem equaon H 1 w. Coupled wh he daa dsrbuon (15), we defne he dynamc generalzed lnear model wh he sae space pror: F v, v ~ N(, V) H 1 w, w ~ N(, W). (25) Jon Poseror Densy Under he usual assumpon ha he pror densy for he covarance marx of srucural and sysem equaons pvw, pv pw, we can express he pror dsrbuon as 1

14 ,,, p V W a T 1,,, 1,, p X V a p W a p V p W. (26) Nex, we se he pror dsrbuon pa of he ndvdual mean parameer a by a N a a N a b IG n s. (27) , / ;, ; /2, /2 By arrangng he erm n (22) as c a, where c log a log x ', for 1,..., T, we consue he lkelhood for he mean a condonal on he error varance and derve condonal poseror dsrbuon p a V W a 2 closed form as,,,,, n a 2 a Tc v, (28) T N, 2 2 T a v where T c c / T and 1 a means he se of ak, k 1,..., I excludng a. Fnally, we oban he jon poseror densy wh daa augmenaon of z, ', ', ', ' ' by equaon (29): * [1] [2] [3] 2 p,, V, W, a, ; z daa * * 3 [ k ] [ k,,,, ] [ k ], T k 1 1 p n p p m n p p y m p I 1,,, 1, p X V a p W p V p W 2 2 p a,, V, W, a, p a p. In (29), daa means he observed daa y, x. (29) Our model wh over-dspersons s characerzed as he dynamc generalzed lnear models 11

15 of he Posson Mulnomal dsrbuon perurbed by he Gamma Drchle dsrbuons. 7. Esmaon and Forecasng Samplng he lnk funcon for MCMC In addon o he sandard Bayesan nference on sae space modelng by dynamc lnear models (DLMs) by Wes and Harrson (1997), we use he MCMC approach o esmae he model by usng Meropols Hasngs samplng specfcally for he condonal poseror densy of lnk funcons based on he daa-augmened represenaon,,,daa,,,, p F V p n m y z p z p F V dz, (3) * [1] [2] [3] where z, ', ', ', ' '. Once he values of for equaon (3) are gven, he srucural equaons coupled wh he sysem equaons n her sae space prors n equaon (25) consue he convenonal Gaussan sae space models. The mul-move sampler by Carer and Kohn (1994) and Fruhwrh-Schnaer (1994) s used o sample he sae vecor. We assume ha he nal values of he sae vecor follow a mulvarae normal dsrbuon N, d I. The mean vecor was se as he esmae of he coeffcen on sac regresson,.e., he regresson wh me-nvaran coeffcen, and we se d =.1 for he emprcal applcaon. Predcve Densy Nex, one-sep-ahead predcve densy p y daa 1 s evaluaed by 12

16 y 1,daa 1 1, ,,,daa T T T T T T T p T 1 T, V, W,daa pv, Wd TdVdW, p z p z p V W (31) where p,,,daa T 1 T 1 V W s he condonal predcve densy of lnk parameers when he predced sae vecor, and srucural and sysem error covarance marces are gven. The MCMC mehod s appled o evaluae hs predcve densy by he augmenng procedure. To evaluae hs densy, we frs exend equaon (16) o defne he predcve lkelhood of condonal on T 1 by T 1,,,,daa Gamma, p z V W a d * * T1 T 1 T1 T1 T1 T1 T1 T1 3 Drchle a Drchle a d d [ k] [ k] [ k] T1 m, T1 T1 T1 T1 T1 k 1 (32) The deals of he samplng scheme for MCMC are descrbed n he appendx. 8. Emprcal Applcaon Daa and Varables We use he sore level scanner, pon of sales (POS), me seres n he curry roux caegory ha was appled o our prevous model n Teru e al. (21) for comparson wh he model wh over-dsperson. The weekly seres comprses hree manufacurers ha produce hree producs each, for a oal of nne producs durng 11 weeks. Table 1 Summary of Daa Fgure 2 Hsogram of Weekly Produc Sales Daa Table 1 descrbes he summary sascs for sales, prce dsplay, and feaure daa. In parcular, sales daa conans varance o gve an evdence of he presence of over-dsperson 13

17 n he daa. Fgure 2 show he hsogram of brand sales. The frs 1 weeks are used for esmaon and he las 1 weeks are reserved for valdaon of forecasng. The daa conan he amoun of produc sales for y, and prces, dsplay (n-sore promoon), and feaures (adversng n newspaper) for markeng mx varables x. x conans no only varables of her own, bu also hose wh oher producs. The dsplay and feaures are bnary daa akng a value of 1 when was on and when was off. The logs of prce daa are used. Model Comparson Each of he hree makers, A, B, and C, produces hree caegores of producs accordng o he level of spcness o accommodae he dfference n consumer ases (1: No spcy, 2: Medum spcy, 3: Spcy). Followng he dscusson of Teru (211), we assume hree possble marke srucures: (1) produc caegory, (2) makers, and (3) usage,.e., ordnary or luxury usage, as shown n Fgure 3, and compare hese models wh over-dsperson and whou over-dsperson. Tha s, we have sx models o be compared. Fgure 3 Comparave Models The op of Table 2-1 shows he log of margnal lkelhood (LML) as an n-sample f creron and wo ypes of predcve measures,.e., he devance nformaon crera (DIC) by Spegelhaler e al. (22), and he roo mean squared errors (RMSE) of 1-sep-ahead forecass of hold-ou samples as ou-of-sample crera. There are hree levels for forecasng he RMSE: marke, submarke, and produc. These errors, ncludng he null model of no srucure, are repored n he lower panel of he able 2-2. We apply wo ypes of measures: sum1 and sum2. sum1 s he sum of all errors 14

18 nduced by he model n whch we have no specfc preference on he levels o be predced. sum2 s defned as he sum of marke and produc errors by consderng ha he numbers of submarkes dffer beween null and oher srucures. Accordng o he hree crera, he proposed models based on compound Posson varables accommodang over-dsperson mprove he models wh no over-dspersons. In parcular, he compound Posson varable model under he marke srucure (3) shows he bes performance. Table 2-1 Model Comparson: Overall Table 2-2 Model Comparson: Decomposon of RMSE Table 3 Esmaes of Produc Inercep Fgure 4 In-sample Performance and Forecasng Marke, Submarke, and Produc Fgure 5 Esmaes of Srucural Parameer The lef panels of Fgure 4 show he predced f of n-sample daa for marke, submarke of usage, and produc levels, where each observaon s denoed by a do and he esmaes are conneced by sragh lnes. We observe ha he model fs he marke sales well over he observaonal perod. The rgh panel of each elemen depcs s 1-sep-ahead forecasng for hese sales, where he mean values of he predced densy a each predcon sep are conneced by a connuous lne, and he 2.5% and 97.5% quanles of he densy a each sep are conneced by dashed lnes. The hold-ou samples are denoed by dos n he fgure. Ths shows ha he marke wll gradually expand over he nex 1 weeks, and hese forecass are conssen wh he movemen of hold-ou samples. We generae he forecass keepng he las observaon x for he predcon seps. T 15

19 Fgure 5(a), (b) show he rajecory of esmaed parameers appears ha and 1,..., 9 a a. I s are flucuang downward for he frs perod and hen urnng upward wh local rends around a mean level of a1,..., a 9 move more heerogeneously wh large flucuaons, whch should be proporonal o he observed dscree oucomes. Fgure 5(c) ndcaes rends for produc A2 and B3 sales, whch belong o a dfferen caegory; shows he oppose rends. Fgure 5(d) depcs he me-varyng prce coeffcen esmaes n response o prce and promoons ( End dsplay and Adversng ). We confrm he compeve relaonshp beween submarkes, and more neresngly, we fnd ha producs B1 and C2 are no hosle o A1 n he sense of prcng sraegy, as hey have he same coeffcen sgn as ha of A1. 9. Concludng Remarks In hs sudy, we proposed a mulvarae me seres model for over-dspersed dscree daa. Theren, we exended he model wh he herarchcal srucure by Teru e al. (21) o accommodae he over-dsperson problem nheren n he modelng of marke srucure based on sales coun dynamcs. We frs dscussed he mechansm of he mcrosrucure for generang over-dsperson n a number of dscree sales daa. The Gamma compound Posson varable for produc sales coun responses and Drchle compound mulnomal varables for produc share are conneced n a herarchcal fashon as a ree srucure for depcng a marke. The model s based on he lkelhood generaed by decomposng sales coun response varables accordng o he degree of compeveness among producs and condonng on her sum, and bulds hem up o hgher levels, represened as a ree srucure. Sae space prors are appled o he lkelhood generaed by he compound dsrbuons o develop dynamc generalzed lnear models for dscree responses wh a 16

20 herarchcal srucure. We frs showed ha over-dsperson s nheren o he problems where consumers are more or less dependen. Then, we model he compound dsrbuons for accommodang over-dspersons. However, nsead of he drec use of he densy funcon of compound dsrbuons, we augmen varables o make he numercal negraons for mxng easer, and provde more effcen algorhms compared o he mehod ha makes drec use of compound dsrbuons. The emprcal analyss by weekly produc sales daa n a sore showed ha our modelng worked well and he models wh over-dsperson, whch s consruced by compound Poson varables, performs beer han he models whou over-dsperson. There are a few problems for fuure research. One s he exenson of he heorecal sudy. The zero-nflaed Posson (ZIP) model by, for example, Lamber (1992) could be also appled o our modelng when he daa conan many zeros. In parcular, hs could be mporan f we furher ncorporae modelng of ndvdual consumer behavor n he analyss. Ths could be accommodaed by mxure dsrbuons hrough herarchcal models, as used n hs sudy. However, hs addonal mxure modelng demands more complcaed compuaon procedures. The expeced gans from hs exenson could no be subsanal, compared wh he developmen of a new model, and hus, we would lke o leave hs modfcaon of he model for fuure research. 17

21 Appendx: MCMC Algorhm A.1 Dynamc Generalzed Lnear Model We summarze he pror and condonal poseror dsrbuon used for our proposed model below. Condonal Poseror Dsrbuon We run 1, MCMC eraons n he model. In all models, we used he las 5, eraons o esmae he poseror dsrbuon of model parameers. When he nal values of he parameers are gven, he condonal poseror densy of he necessary parameers s generaed as follows. () s defned by equaon (23), and we use Meropols Hasngs wh a random walk algorhm,, ( s) ( s1) N 1 I where I s an deny marx wh correspondng dmensons. Accepance probably s defned as ( s) ( s) ( s) p ( ) ( 1) F,, V,daa s s, mn,1. ( s1) ( s1) ( s1) p F,, V,daa Afer obanng he draw of hese parameers, we use Wes and Harrson s (1997) sandard Bayesan nference procedure on sae space modelng by DLM. () The mul-move sampler by Carer and Kohn (1994) and Fruhwrh-Schnaer (1994) s used o sample he sae vecor () * Generae (v) (v) Generae [ k ] Generae (v) a *( s ) * *( s) ( s) from gamma p a, ( s ) ( s) from Drchle p n, a m [ k]( s) [ k [ ] from Drchle ] k ( s p ) a 18

22 Generae a from ( s ) 2 2 b a ma N,, 2 2 b m b m where m a a / m. We se a, b 1 n he emprcal analyss. 1 (v) 2 Generae m ( )/2, /2. 1 2( s ) from IG n 2 m s a a We se n 2, s m n he emprcal analyss. A.2 Forecasng Dscree Oucomes and Consung Predcve Densy Gven he s-h draw of MCMC ( ) ( ) ( ) s, s, s T V W, ( s ) () oban he forecas p,,,daa T 1 T 1 V W p,,,daa T1 T V W from T 1 by he algorhm of Gaussan sae space model; z T from p 1 zt 1 T 1 ( s) * ( s) ( s) [1] ( s) [2] ( s) [3] ( s) zt 1 1, T T 1 ', T 1 ', T 1 ', T 1 ' ' ; () generae he random number ( s ) * ( s n ) T1 Posson T 1 ( s ) ( s ) (v) gven nt ogeher wh he parameer values 1 T 1 ( s ) () oban he forecas samplng from he mulnomal dsrbuon ( s) ( s ) ( s ) ( s ) T 1 T1 T1 T1 o ge he parameer for he marke sales forecas; ( s), generae by mt 1 m n Mulnomal n, for he submarke sales forecass; [ k]( s) (v) gven [ k]( s) ogeher wh he parameer values T 1 m1 of he mulnomal dsrbuon of submarke M k, generae he respecve produc s forecass by samplng from he mulnomal dsrbuon [ k ]( s ) [ k]( s) [ k]( s) [ k ]( s y ) T 1 mt1 Mulnomal m T1, T 1 for k=1,, L; (v) erae seps () (v) M mes. [ k]( s) Then, he emprcal dsrbuon of y 1, s b,..., M approxmaes he predcve [ k ] densy (31) n z y. We se he burn-n parameer b = 5, and he oal number of T1 1 eraons M = 1, for he emprcal applcaon afer checkng he convergence. Seven hours of compuaon were necessary o mplemen our emprcal analyss. By exendng he 19

23 above forecasng seps up o H sep ahead, we obaned he MCMC sample pah [ k ]( s ) [ k ]( s ) [ k ]( s ) 1 2 H y, y,..., y for he jon predcve densy. 2

24 References Bearden,W.O. and Ezel, M.J. (1982), Reference group nfluence on produc and brand purchase decsons, Journal of Consumer Research, vol. 9, pp Cargnon, C., Muller, P. and Wes, M. (1997), Bayesan forecasng of mulnomal me seres Through Condonally Gaussan Dynamc Models, Journal of he Amercan Sascal Assocaon, vol. 92, pp Carer, C.K. and Kohn, R. (1994), On Gbbs samplng for sae space models, Bomerka, vol. 81, pp Fruhwrh-Schnaer, S. (1994), Daa augmenaon and dynamc lnear models, Journal of Tme Seres Analyss, vol. 15, pp Harvey, A.C. and Fernandes, C. (1989), Tme seres models for coun or qualave observaons, Journal of Busness and Economc Sascs, vol. 7, pp Hoadley, B. (1969), The compound mulnomal dsrbuon and bayesan analyss of caegorcal daa from fne populaons, Journal of he Amercan Sascal Assocaon, vol. 64, pp Hyman, H.H. (1942), The psychology of saus. Archves of Psychology, 269, Reprn n H. Hyman & E. Snger (Eds.), Readngs n reference group heory and research (pp ). New York: Free Press, London: Coller-Macmllan Lmed. (Page caons are o he reprn edon). Lamber, D. (1992), Zero-nflaed Posson regresson, wh applcaon o defecs n manufacurng, Technomercs, vol. 34, pp Ord, K., Fernandes, C., and Harvey, A.C. (1993), Tme seres models for mulvarae seres of coun daa, Ed. T. Subba Rao, Developmens n Tme Seres Analyss, pp Park, C.W. and Lessg, V.P. (1977), Sudens and housewves: Dfferences n suscepbly o reference group nfluence, Journal of Consumer Research, vol. 4, pp Spegelhaler, D.J., Bes, N.G., Carln, B.P., and van der Lnde, A. (22), Bayesan measures of model complexy and f, Journal of he Royal Sascal Socey Seres B., vol. 64, pp Teru, N., Ban, M., and Mak, T. (21), Fndng marke srucure by sales coun dynamcs - mulvarae srucural me seres models wh herarchcal srucure for coun daa -, Annals of he Insue of Sascal Mahemacs, vol. 62, pp Ross, P, E, G. Allenby and R. McCulloch (25), Bayesan Sascs n Markeng, John Wley & Sons, New Jersey. Wes, M. and Harrson, P.J. (1997), Bayesan Forecasng and Dynamc Models, 2 nd ed., Sprnger-Verlag, New York. Wes, M., Harrson, P.J., and Mgon, H.S. (1985), Dynamc generalzed lnear models and 21

25 Bayesan forecasng, Journal of he Amercan Sascal Assocaon, vol. 8, pp Wnkelmann, R. (28), Economerc Analyss of Coun Daa (Ffh edon), Sprnger, Hedelberg. 22

26 Table 1 Summary of Daa weekly weekly weekly weekly Brand Sales Prce (/1g) Dsplay Feaure Average (Varance) Average Average Average A (157.47) B ( ) C (234.13) A (2915.8) B2 87. (447.92) C ( ) A (578.82) B ( ) C (959.64) Table 2-1 Model Comparson: Overall ML DIC RMSE(sum1) RMSE(sum2) Posson Null Produc Caegory Maker Usage Compound Posson Null Produc Caegory Maker Usage ML: he log of margnal lkelhood, DIC: Devaon nformaon measure RMSE: he roo mean squared errors of 1 sep ahead forecass 23

27 Table 2-2 Model Comparson: Decomposon of RMSE Posson Null Produc Caegory Maker Usage Marke Marke Marke Marke A Caegory A Maker Usage A A A A A A B B B A C C B Caegory B Maker A B B A B C B B C C B C C Caegory C Maker Usage C A A C B B C C C (sum1) (sum1) (sum1) (sum1) (sum2) (sum2) 48.6 (sum2) (sum2) Null Produc Caegory Compound Posson Maker Usage Marke Marke Marke Marke A Caegory A Maker Usage A A1 2.4 A A A A B B1 3.3 B A C C B Caegory B Maker A B B A B C B B C2 2.5 C B C C Caegory C 51.1 Maker Usage C A A C B B C C C (sum1) (sum1) (sum1) (sum1) (sum2) (sum2) (sum2) 49. (sum2)

28 Table 3 Esmaes of Produc Inercep Caegory Brand Pos Mean Pos S.D. A B Usage 1 C A B C A Usage 2 B C

29 Fgure 1 Marke Srucure (a) No Specfc Srucure n y 1 y 2 y I (b)three-layer Herarchcal Marke Srucure n Marke m1 m2 m3 Submarke (Caegory) [1] y 1 [1] y 2 [1] [2] [2] [2] y N 1 y 1 y 2 [3] [3] [3] y y 1 y 2 y N N 3 2 Produc 26

30 Fgure 2 Hsogram of Weekly Produc Sales Daa A B C A B C A B3 27

31 C3 Fgure 3 Comparave Models Maker Produc Caegory A B C 1 A1 B1 C1 1 2 A2 B2 C2 3 A3 B3 C3 2 Usage Three makers produce hree caegores of producs and hey are classfed no wo groups by her usages. 28

32 Fgure 4 In-sample Performance and Forecasng: Marke, Submarke, and Produc (a) In-sample f Marke Level (b) Forecasng Marke (marke sales) Marke (marke sales) Caegory Level Usage 1 (Caegory sales) Usage 1 (caegory sales) Usage 2 (Caegory sales) Usage 2 (caegory sales) Produc Level Usage 1, A1 (brand sales) Usage 1, A1 (brand sales)

33 Usage 2, B3 (brand sales) Usage 2, B3 (brand sales) Fgure 5 Esmaes of Srucural Parameer (a) Marke, Tau (b) a Usage1 A1 Usage1 B Usage1 C1 Usage1 A

34 Usage1 B2 Usage1 C Usage2 A3 Usage2 B Usage2 C (c) Trend Usage 1, A2, Trend Usage 2, B3, Trend

35 (d) Response Parameers Usage 1, A1, Prce coef Usage 2, B3, Prce coef A1 B1 C1 A2 B2 C2 A3 B3 C3 Usage 1, C2, End coef Usage 1, C2, Ad coef A1 B1 C1 A2 B2 C2 A1 B1 C1 A2 B2 C2 32

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany

John Geweke a and Gianni Amisano b a Departments of Economics and Statistics, University of Iowa, USA b European Central Bank, Frankfurt, Germany Herarchcal Markov Normal Mxure models wh Applcaons o Fnancal Asse Reurns Appendx: Proofs of Theorems and Condonal Poseror Dsrbuons John Geweke a and Gann Amsano b a Deparmens of Economcs and Sascs, Unversy

More information

Bayesian Inference of the GARCH model with Rational Errors

Bayesian Inference of the GARCH model with Rational Errors 0 Inernaonal Conference on Economcs, Busness and Markeng Managemen IPEDR vol.9 (0) (0) IACSIT Press, Sngapore Bayesan Inference of he GARCH model wh Raonal Errors Tesuya Takash + and Tng Tng Chen Hroshma

More information

Solution in semi infinite diffusion couples (error function analysis)

Solution in semi infinite diffusion couples (error function analysis) Soluon n sem nfne dffuson couples (error funcon analyss) Le us consder now he sem nfne dffuson couple of wo blocks wh concenraon of and I means ha, n a A- bnary sysem, s bondng beween wo blocks made of

More information

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6)

Econ107 Applied Econometrics Topic 5: Specification: Choosing Independent Variables (Studenmund, Chapter 6) Econ7 Appled Economercs Topc 5: Specfcaon: Choosng Independen Varables (Sudenmund, Chaper 6 Specfcaon errors ha we wll deal wh: wrong ndependen varable; wrong funconal form. Ths lecure deals wh wrong ndependen

More information

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS

V.Abramov - FURTHER ANALYSIS OF CONFIDENCE INTERVALS FOR LARGE CLIENT/SERVER COMPUTER NETWORKS R&RATA # Vol.) 8, March FURTHER AALYSIS OF COFIDECE ITERVALS FOR LARGE CLIET/SERVER COMPUTER ETWORKS Vyacheslav Abramov School of Mahemacal Scences, Monash Unversy, Buldng 8, Level 4, Clayon Campus, Wellngon

More information

Robustness Experiments with Two Variance Components

Robustness Experiments with Two Variance Components Naonal Insue of Sandards and Technology (NIST) Informaon Technology Laboraory (ITL) Sascal Engneerng Dvson (SED) Robusness Expermens wh Two Varance Componens by Ana Ivelsse Avlés avles@ns.gov Conference

More information

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model

( t) Outline of program: BGC1: Survival and event history analysis Oslo, March-May Recapitulation. The additive regression model BGC1: Survval and even hsory analyss Oslo, March-May 212 Monday May 7h and Tuesday May 8h The addve regresson model Ørnulf Borgan Deparmen of Mahemacs Unversy of Oslo Oulne of program: Recapulaon Counng

More information

January Examinations 2012

January Examinations 2012 Page of 5 EC79 January Examnaons No. of Pages: 5 No. of Quesons: 8 Subjec ECONOMICS (POSTGRADUATE) Tle of Paper EC79 QUANTITATIVE METHODS FOR BUSINESS AND FINANCE Tme Allowed Two Hours ( hours) Insrucons

More information

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA

RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA RELATIONSHIP BETWEEN VOLATILITY AND TRADING VOLUME: THE CASE OF HSI STOCK RETURNS DATA Mchaela Chocholaá Unversy of Economcs Braslava, Slovaka Inroducon (1) one of he characersc feaures of sock reurns

More information

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4

CS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 4 CS434a/54a: Paern Recognon Prof. Olga Veksler Lecure 4 Oulne Normal Random Varable Properes Dscrmnan funcons Why Normal Random Varables? Analycally racable Works well when observaon comes form a corruped

More information

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005

Dynamic Team Decision Theory. EECS 558 Project Shrutivandana Sharma and David Shuman December 10, 2005 Dynamc Team Decson Theory EECS 558 Proec Shruvandana Sharma and Davd Shuman December 0, 005 Oulne Inroducon o Team Decson Theory Decomposon of he Dynamc Team Decson Problem Equvalence of Sac and Dynamc

More information

Fall 2010 Graduate Course on Dynamic Learning

Fall 2010 Graduate Course on Dynamic Learning Fall 200 Graduae Course on Dynamc Learnng Chaper 4: Parcle Flers Sepember 27, 200 Byoung-Tak Zhang School of Compuer Scence and Engneerng & Cognve Scence and Bran Scence Programs Seoul aonal Unversy hp://b.snu.ac.kr/~bzhang/

More information

On One Analytic Method of. Constructing Program Controls

On One Analytic Method of. Constructing Program Controls Appled Mahemacal Scences, Vol. 9, 05, no. 8, 409-407 HIKARI Ld, www.m-hkar.com hp://dx.do.org/0.988/ams.05.54349 On One Analyc Mehod of Consrucng Program Conrols A. N. Kvko, S. V. Chsyakov and Yu. E. Balyna

More information

2. SPATIALLY LAGGED DEPENDENT VARIABLES

2. SPATIALLY LAGGED DEPENDENT VARIABLES 2. SPATIALLY LAGGED DEPENDENT VARIABLES In hs chaper, we descrbe a sascal model ha ncorporaes spaal dependence explcly by addng a spaally lagged dependen varable y on he rgh-hand sde of he regresson equaon.

More information

( ) () we define the interaction representation by the unitary transformation () = ()

( ) () we define the interaction representation by the unitary transformation () = () Hgher Order Perurbaon Theory Mchael Fowler 3/7/6 The neracon Represenaon Recall ha n he frs par of hs course sequence, we dscussed he chrödnger and Hesenberg represenaons of quanum mechancs here n he chrödnger

More information

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction

F-Tests and Analysis of Variance (ANOVA) in the Simple Linear Regression Model. 1. Introduction ECOOMICS 35* -- OTE 9 ECO 35* -- OTE 9 F-Tess and Analyss of Varance (AOVA n he Smple Lnear Regresson Model Inroducon The smple lnear regresson model s gven by he followng populaon regresson equaon, or

More information

Math 128b Project. Jude Yuen

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

More information

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

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

More information

CS286.2 Lecture 14: Quantum de Finetti Theorems II

CS286.2 Lecture 14: Quantum de Finetti Theorems II CS286.2 Lecure 14: Quanum de Fne Theorems II Scrbe: Mara Okounkova 1 Saemen of he heorem Recall he las saemen of he quanum de Fne heorem from he prevous lecure. Theorem 1 Quanum de Fne). Le ρ Dens C 2

More information

( ) [ ] MAP Decision Rule

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

More information

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts

Online Appendix for. Strategic safety stocks in supply chains with evolving forecasts Onlne Appendx for Sraegc safey socs n supply chans wh evolvng forecass Tor Schoenmeyr Sephen C. Graves Opsolar, Inc. 332 Hunwood Avenue Hayward, CA 94544 A. P. Sloan School of Managemen Massachuses Insue

More information

Analysis And Evaluation of Econometric Time Series Models: Dynamic Transfer Function Approach

Analysis And Evaluation of Econometric Time Series Models: Dynamic Transfer Function Approach 1 Appeared n Proceedng of he 62 h Annual Sesson of he SLAAS (2006) pp 96. Analyss And Evaluaon of Economerc Tme Seres Models: Dynamc Transfer Funcon Approach T.M.J.A.COORAY Deparmen of Mahemacs Unversy

More information

Linear Response Theory: The connection between QFT and experiments

Linear Response Theory: The connection between QFT and experiments Phys540.nb 39 3 Lnear Response Theory: The connecon beween QFT and expermens 3.1. Basc conceps and deas Q: ow do we measure he conducvy of a meal? A: we frs nroduce a weak elecrc feld E, and hen measure

More information

Variants of Pegasos. December 11, 2009

Variants of Pegasos. December 11, 2009 Inroducon Varans of Pegasos SooWoong Ryu bshboy@sanford.edu December, 009 Youngsoo Cho yc344@sanford.edu Developng a new SVM algorhm s ongong research opc. Among many exng SVM algorhms, we wll focus on

More information

Advanced time-series analysis (University of Lund, Economic History Department)

Advanced time-series analysis (University of Lund, Economic History Department) Advanced me-seres analss (Unvers of Lund, Economc Hsor Dearmen) 3 Jan-3 Februar and 6-3 March Lecure 4 Economerc echnues for saonar seres : Unvarae sochasc models wh Box- Jenns mehodolog, smle forecasng

More information

Volatility Interpolation

Volatility Interpolation Volaly Inerpolaon Prelmnary Verson March 00 Jesper Andreasen and Bran Huge Danse Mares, Copenhagen wan.daddy@danseban.com brno@danseban.com Elecronc copy avalable a: hp://ssrn.com/absrac=69497 Inro Local

More information

Standard Error of Technical Cost Incorporating Parameter Uncertainty

Standard Error of Technical Cost Incorporating Parameter Uncertainty Sandard rror of echncal Cos Incorporang Parameer Uncerany Chrsopher Moron Insurance Ausrala Group Presened o he Acuares Insue General Insurance Semnar 3 ovember 0 Sydney hs paper has been prepared for

More information

DEEP UNFOLDING FOR MULTICHANNEL SOURCE SEPARATION SUPPLEMENTARY MATERIAL

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

More information

Department of Economics University of Toronto

Department of Economics University of Toronto Deparmen of Economcs Unversy of Torono ECO408F M.A. Economercs Lecure Noes on Heeroskedascy Heeroskedascy o Ths lecure nvolves lookng a modfcaons we need o make o deal wh he regresson model when some of

More information

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION

UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 2017 EXAMINATION INTERNATIONAL TRADE T. J. KEHOE UNIVERSITAT AUTÒNOMA DE BARCELONA MARCH 27 EXAMINATION Please answer wo of he hree quesons. You can consul class noes, workng papers, and arcles whle you are workng on he

More information

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

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

More information

CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING

CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING CHAPTER FOUR REPEATED MEASURES IN TOXICITY TESTING 4. Inroducon The repeaed measures sudy s a very commonly used expermenal desgn n oxcy esng because no only allows one o nvesgae he effecs of he oxcans,

More information

CHAPTER 10: LINEAR DISCRIMINATION

CHAPTER 10: LINEAR DISCRIMINATION CHAPER : LINEAR DISCRIMINAION Dscrmnan-based Classfcaon 3 In classfcaon h K classes (C,C,, C k ) We defned dscrmnan funcon g j (), j=,,,k hen gven an es eample, e chose (predced) s class label as C f g

More information

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS

THE PREDICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS THE PREICTION OF COMPETITIVE ENVIRONMENT IN BUSINESS INTROUCTION The wo dmensonal paral dfferenal equaons of second order can be used for he smulaon of compeve envronmen n busness The arcle presens he

More information

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10)

Fall 2009 Social Sciences 7418 University of Wisconsin-Madison. Problem Set 2 Answers (4) (6) di = D (10) Publc Affars 974 Menze D. Chnn Fall 2009 Socal Scences 7418 Unversy of Wsconsn-Madson Problem Se 2 Answers Due n lecure on Thursday, November 12. " Box n" your answers o he algebrac quesons. 1. Consder

More information

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s

Ordinary Differential Equations in Neuroscience with Matlab examples. Aim 1- Gain understanding of how to set up and solve ODE s Ordnary Dfferenal Equaons n Neuroscence wh Malab eamples. Am - Gan undersandng of how o se up and solve ODE s Am Undersand how o se up an solve a smple eample of he Hebb rule n D Our goal a end of class

More information

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System

Survival Analysis and Reliability. A Note on the Mean Residual Life Function of a Parallel System Communcaons n Sascs Theory and Mehods, 34: 475 484, 2005 Copyrgh Taylor & Francs, Inc. ISSN: 0361-0926 prn/1532-415x onlne DOI: 10.1081/STA-200047430 Survval Analyss and Relably A Noe on he Mean Resdual

More information

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim

GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS. Youngwoo Ahn and Kitae Kim Korean J. Mah. 19 (2011), No. 3, pp. 263 272 GENERATING CERTAIN QUINTIC IRREDUCIBLE POLYNOMIALS OVER FINITE FIELDS Youngwoo Ahn and Kae Km Absrac. In he paper [1], an explc correspondence beween ceran

More information

Clustering (Bishop ch 9)

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

More information

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth

Should Exact Index Numbers have Standard Errors? Theory and Application to Asian Growth Should Exac Index umbers have Sandard Errors? Theory and Applcaon o Asan Growh Rober C. Feensra Marshall B. Rensdorf ovember 003 Proof of Proposon APPEDIX () Frs, we wll derve he convenonal Sao-Vara prce

More information

Graduate Macroeconomics 2 Problem set 5. - Solutions

Graduate Macroeconomics 2 Problem set 5. - Solutions Graduae Macroeconomcs 2 Problem se. - Soluons Queson 1 To answer hs queson we need he frms frs order condons and he equaon ha deermnes he number of frms n equlbrum. The frms frs order condons are: F K

More information

Time Scale Evaluation of Economic Forecasts

Time Scale Evaluation of Economic Forecasts CENTRAL BANK OF CYPRUS EUROSYSTEM WORKING PAPER SERIES Tme Scale Evaluaon of Economc Forecass Anons Mchs February 2014 Worng Paper 2014-01 Cenral Ban of Cyprus Worng Papers presen wor n progress by cenral

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 0 Canoncal Transformaons (Chaper 9) Wha We Dd Las Tme Hamlon s Prncple n he Hamlonan formalsm Dervaon was smple δi δ Addonal end-pon consrans pq H( q, p, ) d 0 δ q ( ) δq ( ) δ

More information

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data

Appendix H: Rarefaction and extrapolation of Hill numbers for incidence data Anne Chao Ncholas J Goell C seh lzabeh L ander K Ma Rober K Colwell and Aaron M llson 03 Rarefacon and erapolaon wh ll numbers: a framewor for samplng and esmaon n speces dversy sudes cology Monographs

More information

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study)

New M-Estimator Objective Function. in Simultaneous Equations Model. (A Comparative Study) Inernaonal Mahemacal Forum, Vol. 8, 3, no., 7 - HIKARI Ld, www.m-hkar.com hp://dx.do.org/.988/mf.3.3488 New M-Esmaor Objecve Funcon n Smulaneous Equaons Model (A Comparave Sudy) Ahmed H. Youssef Professor

More information

Panel Data Regression Models

Panel Data Regression Models Panel Daa Regresson Models Wha s Panel Daa? () Mulple dmensoned Dmensons, e.g., cross-secon and me node-o-node (c) Pongsa Pornchawseskul, Faculy of Economcs, Chulalongkorn Unversy (c) Pongsa Pornchawseskul,

More information

FTCS Solution to the Heat Equation

FTCS Solution to the Heat Equation FTCS Soluon o he Hea Equaon ME 448/548 Noes Gerald Reckenwald Porland Sae Unversy Deparmen of Mechancal Engneerng gerry@pdxedu ME 448/548: FTCS Soluon o he Hea Equaon Overvew Use he forward fne d erence

More information

An introduction to Support Vector Machine

An introduction to Support Vector Machine An nroducon o Suppor Vecor Machne 報告者 : 黃立德 References: Smon Haykn, "Neural Neworks: a comprehensve foundaon, second edon, 999, Chaper 2,6 Nello Chrsann, John Shawe-Tayer, An Inroducon o Suppor Vecor Machnes,

More information

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!") i+1,q - [(!

In the complete model, these slopes are ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL. (! i+1 -! i ) + [(!) i+1,q - [(! ANALYSIS OF VARIANCE FOR THE COMPLETE TWO-WAY MODEL The frs hng o es n wo-way ANOVA: Is here neracon? "No neracon" means: The man effecs model would f. Ths n urn means: In he neracon plo (wh A on he horzonal

More information

A HIERARCHICAL KALMAN FILTER

A HIERARCHICAL KALMAN FILTER A HIERARCHICAL KALMAN FILER Greg aylor aylor Fry Consulng Acuares Level 8, 3 Clarence Sree Sydney NSW Ausrala Professoral Assocae, Cenre for Acuaral Sudes Faculy of Economcs and Commerce Unversy of Melbourne

More information

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

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

More information

Notes on the stability of dynamic systems and the use of Eigen Values.

Notes on the stability of dynamic systems and the use of Eigen Values. Noes on he sabl of dnamc ssems and he use of Egen Values. Source: Macro II course noes, Dr. Davd Bessler s Tme Seres course noes, zarads (999) Ineremporal Macroeconomcs chaper 4 & Techncal ppend, and Hamlon

More information

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes.

J i-1 i. J i i+1. Numerical integration of the diffusion equation (I) Finite difference method. Spatial Discretization. Internal nodes. umercal negraon of he dffuson equaon (I) Fne dfference mehod. Spaal screaon. Inernal nodes. R L V For hermal conducon le s dscree he spaal doman no small fne spans, =,,: Balance of parcles for an nernal

More information

Let s treat the problem of the response of a system to an applied external force. Again,

Let s treat the problem of the response of a system to an applied external force. Again, Page 33 QUANTUM LNEAR RESPONSE FUNCTON Le s rea he problem of he response of a sysem o an appled exernal force. Agan, H() H f () A H + V () Exernal agen acng on nernal varable Hamlonan for equlbrum sysem

More information

TSS = SST + SSE An orthogonal partition of the total SS

TSS = SST + SSE An orthogonal partition of the total SS ANOVA: Topc 4. Orhogonal conrass [ST&D p. 183] H 0 : µ 1 = µ =... = µ H 1 : The mean of a leas one reamen group s dfferen To es hs hypohess, a basc ANOVA allocaes he varaon among reamen means (SST) equally

More information

DYNAMIC ECONOMETRIC MODELS Vol. 8 Nicolaus Copernicus University Toruń 2008

DYNAMIC ECONOMETRIC MODELS Vol. 8 Nicolaus Copernicus University Toruń 2008 DYNAMIC ECONOMETRIC MODELS Vol. 8 Ncolaus Coperncus Unversy Toruń 008 Monka Kośko The Unversy of Compuer Scence and Economcs n Olszyn Mchał Perzak Ncolaus Coperncus Unversy Modelng Fnancal Tme Seres Volaly

More information

Relative controllability of nonlinear systems with delays in control

Relative controllability of nonlinear systems with delays in control Relave conrollably o nonlnear sysems wh delays n conrol Jerzy Klamka Insue o Conrol Engneerng, Slesan Techncal Unversy, 44- Glwce, Poland. phone/ax : 48 32 37227, {jklamka}@a.polsl.glwce.pl Keywor: Conrollably.

More information

Chapter Lagrangian Interpolation

Chapter Lagrangian Interpolation Chaper 5.4 agrangan Inerpolaon Afer readng hs chaper you should be able o:. dere agrangan mehod of nerpolaon. sole problems usng agrangan mehod of nerpolaon and. use agrangan nerpolans o fnd deraes and

More information

Advanced Machine Learning & Perception

Advanced Machine Learning & Perception Advanced Machne Learnng & Percepon Insrucor: Tony Jebara SVM Feaure & Kernel Selecon SVM Eensons Feaure Selecon (Flerng and Wrappng) SVM Feaure Selecon SVM Kernel Selecon SVM Eensons Classfcaon Feaure/Kernel

More information

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that

THEORETICAL AUTOCORRELATIONS. ) if often denoted by γ. Note that THEORETICAL AUTOCORRELATIONS Cov( y, y ) E( y E( y))( y E( y)) ρ = = Var( y) E( y E( y)) =,, L ρ = and Cov( y, y ) s ofen denoed by whle Var( y ) f ofen denoed by γ. Noe ha γ = γ and ρ = ρ and because

More information

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD

HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Journal of Appled Mahemacs and Compuaonal Mechancs 3, (), 45-5 HEAT CONDUCTION PROBLEM IN A TWO-LAYERED HOLLOW CYLINDER BY USING THE GREEN S FUNCTION METHOD Sansław Kukla, Urszula Sedlecka Insue of Mahemacs,

More information

CHAPTER 5: MULTIVARIATE METHODS

CHAPTER 5: MULTIVARIATE METHODS CHAPER 5: MULIVARIAE MEHODS Mulvarae Daa 3 Mulple measuremens (sensors) npus/feaures/arbues: -varae N nsances/observaons/eamples Each row s an eample Each column represens a feaure X a b correspons o he

More information

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015)

5th International Conference on Advanced Design and Manufacturing Engineering (ICADME 2015) 5h Inernaonal onference on Advanced Desgn and Manufacurng Engneerng (IADME 5 The Falure Rae Expermenal Sudy of Specal N Machne Tool hunshan He, a, *, La Pan,b and Bng Hu 3,c,,3 ollege of Mechancal and

More information

Sampling Procedure of the Sum of two Binary Markov Process Realizations

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

More information

Comb Filters. Comb Filters

Comb Filters. Comb Filters The smple flers dscussed so far are characered eher by a sngle passband and/or a sngle sopband There are applcaons where flers wh mulple passbands and sopbands are requred Thecomb fler s an example of

More information

Robust and Accurate Cancer Classification with Gene Expression Profiling

Robust and Accurate Cancer Classification with Gene Expression Profiling Robus and Accurae Cancer Classfcaon wh Gene Expresson Proflng (Compuaonal ysems Bology, 2005) Auhor: Hafeng L, Keshu Zhang, ao Jang Oulne Background LDA (lnear dscrmnan analyss) and small sample sze problem

More information

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading

Online Supplement for Dynamic Multi-Technology. Production-Inventory Problem with Emissions Trading Onlne Supplemen for Dynamc Mul-Technology Producon-Invenory Problem wh Emssons Tradng by We Zhang Zhongsheng Hua Yu Xa and Baofeng Huo Proof of Lemma For any ( qr ) Θ s easy o verfy ha he lnear programmng

More information

PhD/MA Econometrics Examination. January, 2019

PhD/MA Econometrics Examination. January, 2019 Economercs Comprehensve Exam January 2019 Toal Tme: 8 hours MA sudens are requred o answer from A and B. PhD/MA Economercs Examnaon January, 2019 PhD sudens are requred o answer from A, B, and C. The answers

More information

Lecture 11 SVM cont

Lecture 11 SVM cont Lecure SVM con. 0 008 Wha we have done so far We have esalshed ha we wan o fnd a lnear decson oundary whose margn s he larges We know how o measure he margn of a lnear decson oundary Tha s: he mnmum geomerc

More information

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance

Bayes rule for a classification problem INF Discriminant functions for the normal density. Euclidean distance. Mahalanobis distance INF 43 3.. Repeon Anne Solberg (anne@f.uo.no Bayes rule for a classfcaon problem Suppose we have J, =,...J classes. s he class label for a pxel, and x s he observed feaure vecor. We can use Bayes rule

More information

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5

[ ] 2. [ ]3 + (Δx i + Δx i 1 ) / 2. Δx i-1 Δx i Δx i+1. TPG4160 Reservoir Simulation 2018 Lecture note 3. page 1 of 5 TPG460 Reservor Smulaon 08 page of 5 DISCRETIZATIO OF THE FOW EQUATIOS As we already have seen, fne dfference appromaons of he paral dervaves appearng n he flow equaons may be obaned from Taylor seres

More information

Machine Learning Linear Regression

Machine Learning Linear Regression Machne Learnng Lnear Regresson Lesson 3 Lnear Regresson Bascs of Regresson Leas Squares esmaon Polynomal Regresson Bass funcons Regresson model Regularzed Regresson Sascal Regresson Mamum Lkelhood (ML)

More information

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. number of linearly independent eigenvectors associated with this eigenvalue. Lnear Algebra Lecure # Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons

More information

Advanced Macroeconomics II: Exchange economy

Advanced Macroeconomics II: Exchange economy Advanced Macroeconomcs II: Exchange economy Krzyszof Makarsk 1 Smple deermnsc dynamc model. 1.1 Inroducon Inroducon Smple deermnsc dynamc model. Defnons of equlbrum: Arrow-Debreu Sequenal Recursve Equvalence

More information

Kayode Ayinde Department of Pure and Applied Mathematics, Ladoke Akintola University of Technology P. M. B. 4000, Ogbomoso, Oyo State, Nigeria

Kayode Ayinde Department of Pure and Applied Mathematics, Ladoke Akintola University of Technology P. M. B. 4000, Ogbomoso, Oyo State, Nigeria Journal of Mahemacs and Sascs 3 (4): 96-, 7 ISSN 549-3644 7 Scence Publcaons A Comparave Sudy of he Performances of he OLS and some GLS Esmaors when Sochasc egressors are boh Collnear and Correlaed wh

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm H ( q, p, ) = q p L( q, q, ) H p = q H q = p H = L Equvalen o Lagrangan formalsm Smpler, bu

More information

Lecture Notes 4. Univariate Forecasting and the Time Series Properties of Dynamic Economic Models

Lecture Notes 4. Univariate Forecasting and the Time Series Properties of Dynamic Economic Models Tme Seres Seven N. Durlauf Unversy of Wsconsn Lecure Noes 4. Unvarae Forecasng and he Tme Seres Properes of Dynamc Economc Models Ths se of noes presens does hree hngs. Frs, formulas are developed o descrbe

More information

Vegetable Price Prediction Using Atypical Web-Search Data

Vegetable Price Prediction Using Atypical Web-Search Data Vegeable Prce Predcon Usng Aypcal Web-Search Daa Do-l Yoo Deparmen of Agrculural Economcs Chungbuk Naonal Unversy Emal: d1yoo@chungbuk.ac.kr Seleced Paper prepared for presenaon a he 2016 Agrculural &

More information

Normal Random Variable and its discriminant functions

Normal Random Variable and its discriminant functions Noral Rando Varable and s dscrnan funcons Oulne Noral Rando Varable Properes Dscrnan funcons Why Noral Rando Varables? Analycally racable Works well when observaon coes for a corruped snle prooype 3 The

More information

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations

[Link to MIT-Lab 6P.1 goes here.] After completing the lab, fill in the following blanks: Numerical. Simulation s Calculations Chaper 6: Ordnary Leas Squares Esmaon Procedure he Properes Chaper 6 Oulne Cln s Assgnmen: Assess he Effec of Sudyng on Quz Scores Revew o Regresson Model o Ordnary Leas Squares () Esmaon Procedure o he

More information

Introduction to Boosting

Introduction to Boosting Inroducon o Boosng Cynha Rudn PACM, Prnceon Unversy Advsors Ingrd Daubeches and Rober Schapre Say you have a daabase of news arcles, +, +, -, -, +, +, -, -, +, +, -, -, +, +, -, + where arcles are labeled

More information

ABSTRACT KEYWORDS. Bonus-malus systems, frequency component, severity component. 1. INTRODUCTION

ABSTRACT KEYWORDS. Bonus-malus systems, frequency component, severity component. 1. INTRODUCTION EERAIED BU-MAU YTEM ITH A FREQUECY AD A EVERITY CMET A IDIVIDUA BAI I AUTMBIE IURACE* BY RAHIM MAHMUDVAD AD HEI HAAI ABTRACT Frangos and Vronos (2001) proposed an opmal bonus-malus sysems wh a frequency

More information

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy

Approximate Analytic Solution of (2+1) - Dimensional Zakharov-Kuznetsov(Zk) Equations Using Homotopy Arcle Inernaonal Journal of Modern Mahemacal Scences, 4, (): - Inernaonal Journal of Modern Mahemacal Scences Journal homepage: www.modernscenfcpress.com/journals/jmms.aspx ISSN: 66-86X Florda, USA Approxmae

More information

Mechanics Physics 151

Mechanics Physics 151 Mechancs Physcs 5 Lecure 9 Hamlonan Equaons of Moon (Chaper 8) Wha We Dd Las Tme Consruced Hamlonan formalsm Hqp (,,) = qp Lqq (,,) H p = q H q = p H L = Equvalen o Lagrangan formalsm Smpler, bu wce as

More information

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION

A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION S19 A NEW TECHNIQUE FOR SOLVING THE 1-D BURGERS EQUATION by Xaojun YANG a,b, Yugu YANG a*, Carlo CATTANI c, and Mngzheng ZHU b a Sae Key Laboraory for Geomechancs and Deep Underground Engneerng, Chna Unversy

More information

Epistemic Game Theory: Online Appendix

Epistemic Game Theory: Online Appendix Epsemc Game Theory: Onlne Appendx Edde Dekel Lucano Pomao Marcano Snscalch July 18, 2014 Prelmnares Fx a fne ype srucure T I, S, T, β I and a probably µ S T. Le T µ I, S, T µ, βµ I be a ype srucure ha

More information

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current :

10. A.C CIRCUITS. Theoretically current grows to maximum value after infinite time. But practically it grows to maximum after 5τ. Decay of current : . A. IUITS Synopss : GOWTH OF UNT IN IUIT : d. When swch S s closed a =; = d. A me, curren = e 3. The consan / has dmensons of me and s called he nducve me consan ( τ ) of he crcu. 4. = τ; =.63, n one

More information

Part II CONTINUOUS TIME STOCHASTIC PROCESSES

Part II CONTINUOUS TIME STOCHASTIC PROCESSES Par II CONTINUOUS TIME STOCHASTIC PROCESSES 4 Chaper 4 For an advanced analyss of he properes of he Wener process, see: Revus D and Yor M: Connuous marngales and Brownan Moon Karazas I and Shreve S E:

More information

Time-interval analysis of β decay. V. Horvat and J. C. Hardy

Time-interval analysis of β decay. V. Horvat and J. C. Hardy Tme-nerval analyss of β decay V. Horva and J. C. Hardy Work on he even analyss of β decay [1] connued and resuled n he developmen of a novel mehod of bea-decay me-nerval analyss ha produces hghly accurae

More information

2 Aggregate demand in partial equilibrium static framework

2 Aggregate demand in partial equilibrium static framework Unversy of Mnnesoa 8107 Macroeconomc Theory, Sprng 2009, Mn 1 Fabrzo Perr Lecure 1. Aggregaon 1 Inroducon Probably so far n he macro sequence you have deal drecly wh represenave consumers and represenave

More information

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue.

. The geometric multiplicity is dim[ker( λi. A )], i.e. the number of linearly independent eigenvectors associated with this eigenvalue. Mah E-b Lecure #0 Noes We connue wh he dscusson of egenvalues, egenvecors, and dagonalzably of marces We wan o know, n parcular wha condons wll assure ha a marx can be dagonalzed and wha he obsrucons are

More information

Existence and Uniqueness Results for Random Impulsive Integro-Differential Equation

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

More information

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective

Forecasting customer behaviour in a multi-service financial organisation: a profitability perspective Forecasng cusomer behavour n a mul-servce fnancal organsaon: a profably perspecve A. Audzeyeva, Unversy of Leeds & Naonal Ausrala Group Europe, UK B. Summers, Unversy of Leeds, UK K.R. Schenk-Hoppé, Unversy

More information

Lecture 2 M/G/1 queues. M/G/1-queue

Lecture 2 M/G/1 queues. M/G/1-queue Lecure M/G/ queues M/G/-queue Posson arrval process Arbrary servce me dsrbuon Sngle server To deermne he sae of he sysem a me, we mus now The number of cusomers n he sysems N() Tme ha he cusomer currenly

More information

P R = P 0. The system is shown on the next figure:

P R = P 0. The system is shown on the next figure: TPG460 Reservor Smulaon 08 page of INTRODUCTION TO RESERVOIR SIMULATION Analycal and numercal soluons of smple one-dmensonal, one-phase flow equaons As an nroducon o reservor smulaon, we wll revew he smples

More information

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management

NPTEL Project. Econometric Modelling. Module23: Granger Causality Test. Lecture35: Granger Causality Test. Vinod Gupta School of Management P age NPTEL Proec Economerc Modellng Vnod Gua School of Managemen Module23: Granger Causaly Tes Lecure35: Granger Causaly Tes Rudra P. Pradhan Vnod Gua School of Managemen Indan Insue of Technology Kharagur,

More information

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

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

More information

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria

M. Y. Adamu Mathematical Sciences Programme, AbubakarTafawaBalewa University, Bauchi, Nigeria IOSR Journal of Mahemacs (IOSR-JM e-issn: 78-578, p-issn: 9-765X. Volume 0, Issue 4 Ver. IV (Jul-Aug. 04, PP 40-44 Mulple SolonSoluons for a (+-dmensonalhroa-sasuma shallow waer wave equaon UsngPanlevé-Bӓclund

More information

Machine Learning 2nd Edition

Machine Learning 2nd Edition INTRODUCTION TO Lecure Sldes for Machne Learnng nd Edon ETHEM ALPAYDIN, modfed by Leonardo Bobadlla and some pars from hp://www.cs.au.ac.l/~aparzn/machnelearnng/ The MIT Press, 00 alpaydn@boun.edu.r hp://www.cmpe.boun.edu.r/~ehem/mle

More information

Dynamic Team Decision Theory

Dynamic Team Decision Theory Dynamc Team Decson Theory EECS 558 Proec Repor Shruvandana Sharma and Davd Shuman December, 005 I. Inroducon Whle he sochasc conrol problem feaures one decson maker acng over me, many complex conrolled

More information