Loss Distribution Generation in Credit Portfolio Modeling

Size: px
Start display at page:

Download "Loss Distribution Generation in Credit Portfolio Modeling"

Transcription

1 Loss Dstrbuto Geerato Credt Portfolo Modelg Igor Jouravlev, MMF, Walde Uversty, USA Ruth A. Maurer, Ph.D., Professor Emertus of Mathematcal ad Computer Sceces, Colorado School of Mes, USA Key words: Loss dstrbuto, Copula, Credt Portfolo, Large Homogeeous Portfolo, Mote Carlo Smulatos, Importace Samplg, Saddle-pot Approxmato Itroducto I the curret paper we aalyze several methods for geerato of loss dstrbuto for credt portfolos. Loss dstrbutos play a mportat role prcg of credt dervatves ad credt portfolo optmzato. A loss dstrbuto s a fucto of the umber of ettes the portfolo, ther credt ratgs, the otoal amout ad recovery of each etty, default probabltes, loss gve defaults, ad the correlato/depedece structure betwee ettes corporated the portfolo. Drect geerato of loss dstrbuto may requre Mote Carlo smulato whch s tme cosumg ad s ot effectve whe appled for credt portfolo optmzato. To overcome computatoal complexty a umber of approaches were udertae based o assumptos mposed o the put parameters, goals of loss dstrbutos geerato, ad the accepted level of tolerace ad computatoal errors. Lterature revew A wde rage of lterature was dedcated to geerato of loss dstrbutos for credt portfolos. Vasce (987, 00) developed a large homogeeous portfolo approxmato that played a mportat role oe of the frst sythetc CDOs prcg methods developed by J.P. Morga. Ths method was geeralzed by cosderg a fte umber of oblgors the portfolo. Hull ad Whte (004) troduced a bucetg approach, ad Aderse, Sdeus ad Basu

2 (003) offered a recursve method vald more geeral cases. Glasserma ad L (005) proposed the two-step-mportace samplg method whch they appled Mote Carlo smulato methods wth varace reducto techques. A Fourer aalytcal approach loss dstrbuto geerato was aalyzed by Mero ad Nyfeler (00), Ress (003), ad Grude (007) whch was used for prcg of CDO by Greogry ad Lauret (004), Lauret (004), ad Lauret ad Gregory (005). Ths method requres a good mplemetato of the fast Fourer trasform. Saddle-pot approxmato was aalyzed by Arvats ad Gregory (00), Gordy (00), Mart (006), Glasserma (008). We wll aalyze ad compare these methods, ther advatages ad dsadvatages. Large homogeeous approxmatos of loss dstrbutos Large homogeeous portfolo (LHP) approxmato was the frst method developed by Oldrch Vasce 987 at KMV Corporato. I ths model, Vasce assumed that the portfolo cotas a fte umber of ettes. Each etty has the same otoal amout, default probablty, ad recovery rate (or loss gve default). Loss gve default ca be calculated as recovery rate. The correlato structure s preseted usg oe-factor Gaussa copula. Although these assumptos were very restrctve, the closed formula derved by Vasce was easy to mplemet ad the model was much faster tha the oe used Mote Carlo smulatos. Accordg to ths model, the rs eutral portfolo cumulatve loss dstrbuto ad probablty desty fucto for a large portfolo wth uderlyg Gaussa copula ca be expressed as follows (Vasce, 00): ρ Φ ( x) Φ ( p) P ( L x) = Φ () ρ

3 3 f ρ ( x) = exp ( ρφ ( x) Φ ( p) ) + ( Φ ( x) ) ρ ρ () where L [0,] portfolo loss estmated as a fracto of the whole portfolo, Φ cumulatve ormal dstrbuto, Φ s the verse of cumulatve ormal dstrbuto, p the probablty of default of a loa the portfolo; ths case we assume that all credts have the same probablty of default, ad the umber of credts s fte, ρ - correlato corporated to the credt portfolo usg oe-factor Gaussa copula. Although the portfolo cotas loas wth the same default characterstcs; the loas are fact dfferet, ad, moreover, correlated wth correlato ρ. The advatage of usg ths formula s that t ca be used for fast approxmato of loss dstrbuto for ay rage [ a, b], where a < b, ad a, b [0,]. The above metoed formulas () ad () deped o very mportat parameters correlato ad probablty of default of a loa the portfolo. A credt portfolo wth hgh value of correlato (say 90%) wth very small value of probablty of default of a loa (say < 0 bps) wll have loss dstrbuto cocetrated aroud 0%, ad ca be approxmately cosdered as a credt portfolo rsless. Oe of the extesos of the Gaussa LHP approach s to use double-t oe factor model proposed by Hull ad Whte (004). They assumed that commo ad dvdual factors are t- dstrbuted ad derved a formula that gves fast approxmato of the loss cumulatve dstrbuto fucto. The LHP approxmato ca be exteded to the case of the Studet-t copula. Ths approach also allows oe to obta aalytcal formulas for desty ad the cumulatve dstrbuto fucto of the portfolo loss dstrbuto (Schloegl & O'Kae, 005). Schloegl ad

4 4 O Kae (005) also compared the VaR mpled by the Studet-t copula to that obtaed usg the Gaussa, Calyto, ad Gumbel copulas. Accordg to Schloegl ad O Kae (005), the returs ζ of each oblgor follow a multvarate Studet-t dstrbuto, so that (p. 578) ζ βz + β ε = (3) W ν ζ where Z, ε,..., ε ~ N(0,), W ~ χ ( ν ), β M ]0,[. Issuer defaults f ad oly f D where D s a certa threshold. Ths equalty s equvalet to: Codtoally o, the default dcator fuctos W β ε D βz =. (4) ν ζ ] are all depedet ad the codtoal default probablty ca be wrtte as a fucto of the stadard ormal cumulatve desty fucto [ ] P ζ D = Φ. By the law of large umbers, the fracto of β [ D defaultg ssuers wll ted to Φ β for each realzato of as M. Assumg that exactly ths fracto of ssuer defaults for each realzato of, thereby replacg the radom varable L by [ L ] E, we have that h( ) [ ] ( ) ( θ [0,] ) P θ = F ( θ ) L, where h( x = ( R) x ) Φ, ad β h, where F s the cumulatve dstrbuto fucto of. Based o ths logc, aalytc formulas for cumulatve dstrbuto fucto ad probablty desty fucto ca be expressed usg the followg formulas (Schloegl & O Kae (005), pp ):

5 5 t u t ν ν ( t + βu) F( t) P[ t], e du D = = Φ + β Γ (5) β π f ( t) = β π ν + ν Γ w t D w + e ν w β ν e 0 dw. (6) As was oted by O Kae (008), there are approaches that approxmately estmate loss dstrbutos ad ca serve as acceptable compromses to the trade-off betwee speed ad accuracy. These approaches are the Gaussa approxmato, bomal ad adjusted bomal dstrbutos (pp ). The Gaussa approxmato uses a Gaussa desty whch fts the frst two momets of the codtoal loss dstrbuto. It s the possble to obta a closed-form expresso for the expected trache loss codtoal o the maret factor (O Kae, 008, p. 354). The dea behd the bomal approxmato s to approxmate the exact multomal dstrbuto wth a bomal dstrbuto. The reaso for ths s that the shape of bomal dstrbuto s a better ft to the multomal dstrbuto tha Gaussa. However, ths approach we match the frst momet of the exact codtoal loss dstrbuto. The further mprovemet s based fdg a way to ft the varace. For ths purpose, a adjusted bomal approxmato was proposed by O Kae (008) to esure that we match frst two momets of the exact homogeeous loss dstrbutos (pp ). The LHP approxmato ca be exteded by usg ormal verse Gaussa dstrbuto (NIG). The ormal verse Gaussa dstrbuto s a mxture of ormal ad verse Gaussa dstrbuto. Accordg to Kalemaova, Schmd, ad Werer (007), a o-egatve radom varable Y has verse Gaussa dstrbuto wth postve parameters α ad β f ts desty fucto ca be represeted usg the followg form:

6 6 f IG α 3 y e ( y; α, β ) = πβ 0, otherwse ( α βy ) βy,f y > 0 (7) A radom varable X follows the NIG dstrbuto wth parameters α, β, μ, δ ( X ~ NIG( α, β, μ, δ ) ) f X Y = y ~ N( μ + βy, y) Y ~ IG( δγ, γ ), γ = α β (8) formula: where The parameters should satsfy the followg codtos: 0 β < α ad δ > 0. The desty of the radom varable X ~ NIG( α, β, μ, δ ) s gve by the followg ( δγ + β ( x μ) ) + ( x μ) αδ exp ( ) f = + ( ) NIG x; α, β, μ, δ K α δ x μ (9) π δ K 0 ( w) = exp w( t + t ) dt (0) s the modfed Bessel fucto of the thrd d. Kalemaova et al. (007) suggested usg the followg parameters the LHP model wth NIG copula (wth the depedecy parameter defed as a square root of correlato assumed a credt portfolo): 3 βγ γ M ~ NIG α, β,, α α V ~ NIG α, β, βγ α, 3 γ α () where M s a systematc maret rs factor, ad are dosycratc factors. V The, asset returs wll also follow NIG dstrbuto wth the followg parameters:

7 7 3,,, ~ α γ α βγ β α NIG M () The thrd ad fourth parameters were chose to get expected value of zero ad varace of. Usg the followg otato = 3 ) (,,, ; ) ( α γ α βγ β α s s s s x F x F NIG s NIG, the default threshold ca be computed as follows: ) ( p F C NIG = ; (3) the default probablty of each credt codtoal o maret factor s gve by = ) ( M C f M p NIG (4) ad the loss dstrbuto of the LHP ca be estmated usg the followg formula (Kalemaova et al., 007): = ) ( ) ( () x F C F x F NIG NIG (6) These formulas use specal fuctos; however, the advatage of usg them s that computato of loss dstrbutos usg Gaussa copula, Studet-t, double-t, or NIG copula s much less tme cosumg tha usg Mote Carlo smulato for geerato loss dstrbutos. Oe of the advatages of usg NIG copula s the possblty of estmatg four parameters of ths dstrbuto gve observed frst four momets. Schöbucher (00) used a algorthm from the theory of Archmedea copula fuctos to estmate lmtg loss dstrbutos whch are drve by radom varable wth dfferet

8 8 depedecy structures. Ths approach allowed presetg smple ad realstc formulas for the loa portfolo dstrbuto. The jot dstrbutos a credt portfolo are modeled dfferet ways tha just usg a varat of the multvarate ormal dstrbuto fucto, ad ths approach proved to be feasble. I obtag loss dstrbutos, t s mportat to vestgate the effect of the mplct assumpto of a Gaussa depedecy structure o the rs measures ad the returs dstrbuto of the portfolo, as well as the cosequeces of extreme evets ad lac of avalable data o credt rs modelg. Schöbucher (00) showed that the credt rs case, ths effect ca be ether mor (whe the Vasce model s compared to the Clayto-depedet model) or sgfcat (whe oe ths that the Gumbel copula s a realstc alteratve). Fte homogeeous approxmato of loss dstrbutos I fte homogeeous approxmato (ofte called exact computato), the fty assumpto s dropped ad the model uses assumptos of a sgle systematc factor ad homogeety; ths case the umercal procedure s stll easy to mplemet. I most cases, however, the portfolos are ot homogeeous, but f we assume that the portfolo s large, graularty adjustmet developed by Gordy (003), Pyht ad Dev (003), ad Gordy (004) ca be appled. I both large homogeeous ad fte homogeeous portfolos the loss dstrbuto ca be geerated based o assumpto of several systematc rs factors usg approprate radom umber geeratos for each factors. For a credt portfolo cosstg of ettes, the Gaussa copula case, the probablty of defaults (or ucodtoal loss dstrbuto dscrete case) ca be expressed usg the followg formula (Vasce, 00):

9 9 P = Φ ρ ( ( p) u) Φ Φ ρ Φ ( p) ρ ( ρu) dφ( u) (7) where p s the probablty of default, ρ s correlato corporated to the credt portfolo usg Gaussa copula, the tegrad s the codtoal probablty of the portfolo loss gve the maret factor u whch s assumed to be ormally dstrbuted. If we cosder m maret factors ths approach, the the tegrato would be over these m maret factors. Codtoal ad ucodtoal loss dstrbutos geerato The prevous formula developed by Vasce (00) gves a dea o how to obta the ucodtoal loss dstrbuto the geeral case. Frst of all, oe has to compute a codtoal loss dstrbuto codtoal o a set of uderlyg factors whch defaults are depedet, ad the tegrate the codtoal loss dstrbuto over the dstrbuto of the uderlyg factors. I mathematcal otato, we eed to compute codtoal probabltes codtoal o maret factors frst: M M M [ ] ( ) (,,..., l = L M, M,..., M = p P L M, M,..., M l l t pt ) (8) ad the tegrate over these factors: where L ( R [ L L M = m,..., M l = ml ] df( m ) df( ml )... P =... (9) = )s the percetage portfolo loss, P [ L L M, M,..., ] = s the probablty M l that exactly out of ssuers default codtoal o maret factors M M,..., ; ad M M p,..., l t, s the codtoal default probablty of oblgor at tme t. I case of Gaussa copula ad case whe we cosder oly oe maret factor our model, we have: M l

10 0 p M t Φ = Φ ( p ) ρ M. (0) ρ Edgeworth expaso ca be used as oe of the methods for geeratg of loss dstrbuto (Arvats & Gregory, 00). Ths expaso uses the hgher-order momets of the dstrbutos of the costtuet varables (such as umber of defaults) ad formato cotaed cumulats. It s well ow that for the ormal dstrbuto, the frst two cumulats are the mea ad varace ad the others are equal to zero. The hgher cumulats gve quattatve formato about the o-ormalty of a dstrbuto. Edgeworth expaso states that f the umber of defaults are depedet radom varables wth meas p, stadard devato σ ad cumulats κ, the probablty desty fucto of the radom varable Y = d = whch represets umber of defaults, ca be gve by f Y ( y) dy = t / e dt κ 3H 3( t) + 3 π 3! σ κ 4H 4 ( t) κ 3 H 6 ( t) ! σ!3! σ +... () wth y p t =, p = p, σ = σ, κ r = κ, σ = = = r () where H r (t) s the rth Hermte polyomal, whch ca be obtaed by successve dfferetato of the fucto t e usg followg equato: t t r d H r ( t) = ( ) e e (3) r dt Cumulats ca be obtaed from the power-seres expaso of the logarthm of the momet geeratg fucto of the radom varable. The leadg term ths expaso s the

11 ormal dstrbuto, ad the frst ad secod terms adjust the sewess ad urtoss. These adjustmets gve better approxmato to the loss dstrbuto by corporatg of the sew (whch correspods to the asymmetry of the dstrbuto) ad urtoss (whch correspods to the fat tals of dstrbuto). As was show by Arvats ad Gregory (00), the Edgeworth approxmato wth four momets s better tha the ormal approxmato, but t becomes less accurate further to the tal. The approxmato s poor to the left of the org, where true probablty desty fucto vashes. For example, the estmato of the uexpected loss at the 99.9 th percetle correspods to a tal probablty of The ormal approxmato uderestmates the true value by 40% ad the Edgeworth approxmato (wth four terms) uderestmates t by 4% (pp ). Hull ad Whte (004) preseted two approaches geeratg loss dstrbutos. They cosdered a umber of maret factors M,..., M l cosdered codtoal o these maret factors. Defg ad the codtoal default probabltes were π T () the probablty that exactly defaults occur the portfolo before tme T, codtoal o the default tmes t are depedet. The codtoal default probablty that all the ames wll survve beyod tme T s π T ( 0 V,..., Vm ) = S ( T M,..., M l ) where S ( T M,..., M l ) s the survval probablty of the = oblgor. Hull ad Whte (004) showed that the codtoal probablty of exactly defaults by tme T s T ( M,..., M l ) π T ( 0 M,..., M l ) π (4) = wz() wz()... wz( )

12 where w ( T M,..., M l ) ( T M,..., M ) S = ad z( ),... z() s the set of dfferet umbers from the fte S l set{,,}. Hull ad Whte (004) provded a fast algorthm for computg the codtoal losses. By tegratg over the maret factors, oe ca obta ucodtoal loss dstrbuto. Hull ad Whte (004) proposes also a bucetg approach whle Aderso et al. (003) proposed a recursve method for geeratg loss dstrbuto. I both approaches, loss gve default ad otoal values ca vary betwee the ettes, but they are stll assumed to be determstc. Hull ad Whte (004) dvded potetal losses to the followg rages: { 0 }{ } {, b, b, b,..., b, 0 0 K b } where { b, b } s referred as th bucet. The loss dstrbuto s bult by cludg oe debt strumet at a tme. The procedure eeps trac of both the probablty of the cumulatve loss beg a bucet ad the mea cumulatve loss codtoal that the cumulatve loss s the bucet. The approach offered by Hull ad Whte (004) does ot assume bucets of detcal sze; t allows the aalyst accommodatg stuatos where extra accuracy s eeded some regos of the loss dstrbuto. The latter ca be acheved by cosderg smaller bucet szes. The loss dstrbuto ca be trucated at some level so that the aalyst eed ot sped extra computatoal tme o large losses that have oly a very small chace of occurrg. Hull ad Whte (004) reported that ther approach s comparable wth the Fourer-aalytcal approach terms of computatoal tme ad accuracy ad s umercally stable. The ma dea of the approach suggested by Aderse et al. (003) was to compute some u as a commo dvsor of all potetal losses, ad the cosder losses * 0, u,u,..., u rouded to the earest dscrete pot as the loss dstrbuto s bult up. * u here s the maxmum possble

13 3 loss. It s mportat to ote here that the speed of ths algorthm depeds o the statstcal characterstcs of the credt spreads, so that some cases, the value of the commo dvsor of all potetal losses ca be very small substatally affectg the speed of the algorthm. The ext step was to mplemet a recursve algorthm to determe the portfolo loss dstrbuto that s used for the codtoal probabltes ad case whe default evets are depedet. It should be oted here that the value for u ca be very small some credt portfolos ad cur computatoally extesve loss dstrbuto geerato. Suppose we ow the loss dstrbuto P K ( l; t), l = 0,..., l max, K for a referece pool of some sze K 0, where s the sum of all the loss weghts such referece pool. Suppose we add aother compay to the pool wth loss weght w K + ad ow default probablty p K loss dstrbuto of the larger baset (Aderse et al., 003, p.67): ( t) l max, K +. The usg depedece of defaults we fd for the K ( pk + ( t) ) + P ( l wk + ; t) p K + ( t), l = max, K + w K + K P ( l; t) = P ( l; t)! 0,..., l + (5) Ths recursve relato starts wth a empty baset, ad crease baset sze leads to the same relatve crease the maxmal loss. The cost of buldg the codtoal loss dstrbuto grows as roughly the square of the baset sze (Aderse et al., 003, p.67). The resultg codtoal loss dstrbuto s trasformed to the ucodtoal loss dstrbuto by tegratg over commo factor. Fourer aalytcal approach s aother approach of codtoal ad ucodtoal loss dstrbuto geerato. Ths approach s alteratve to recurso techques ad cosders a map of the orgal problem to aother space where the problem s more aalytcally tractable. Oce the problem ths space s solved, we eed to map the soluto bac to the orgal space. Ths approach depeds o the successful mplemetato of fast Fourer techques. Ths approach was used, for example, by Gregory ad Lauret (003, 004) for prcg CDOs. Reβ (003)

14 4 provded detaled formato o how loss dstrbuto ad ts frst momets ca be obtaed usg Fourer aalytcal approach, descrbed the CredtRs+ model terms of characterstc fuctos stead of probablty fuctos. Sce costructo of loss dstrbuto very much depeds o the basc loss ut, Reβ preseted a alteratve approach where o basc loss ut had to be troduced. Mero ad Nyfeler (00) descrbed a algorthm that combes techques from umercal mathematcs ad actuaral scece. Ther approach geerato of loss dstrbutos was groupg all potetal losses to exposure bucets. Approxmatg the Beroull default dcators by Posso radom varables allowed reducg the umber of radom varables whch correspod to the umber of credts the portfolo to the umber of bucets. Applyg the fast Fourer trasform, umercal quas Mote Carlo methods allowed geeratg loss dstrbutos of credt portfolos cotag 500,000 couterpartes wth four hours wth adequate accuracy. It was show that t was ot ecessary to smplfy the credt rs model or portfolo structure to calculate the body ad the tal of the portfolo loss dstrbuto ad that the algorthm was useful for aalyzg ad desgg CDO structures. Grude (007) aalyzed whether a Fourer-based approach could be a effcet for calculato of rs measures the cotext of a credt portfolo model wth tegrated maret rs factors. He appled ths approach to CredtMetrcs credt portfolo model exteded by correlated terest rate ad credt spread rs. He showed that Fourer-base methods beg superor to Mote-Carlo smulatos could t be superor case of the tegrated maret ad credt portfolo model eve after applyg stadard mportace samplg techques for mprovg the performace of Fourer-based approach. Ths s because the hgher the cofdece level of the VaR, the larger the asset retur correlato, or the larger the umber of systematc rs factors. For the tegrated maret ad credt portfolo methods, oe should combe the Mote Carlo

15 5 smulato wth a mportace samplg techque. Combato of Mote-Carlo ad mportace samplg techques would be approprate for those cases where the Fourer-based approach performs badly, for example, for the estmato of small percetles whch are eeded credt rs maagemet. The FFT approach was the frst used loss dstrbuto costructo methodology. However, t s slower approach compared to the recursve approach (but stll faster tha the Mote Carlo approach) of geeratg loss dstrbutos due to the followg reasos (O Kae, 008): recursos are faster tha Fourer methods, recursos are easer to mplemet ad do t requre access to ay specalzed umercal lbrares; usg recursos the researcher ca buld the loss dstrbuto for a specfc trache; ths s ot possble to do whe usg FFT. The saddle-pot approxmato for geeratg loss dstrbuto proved to be very accurate practce. Whe cosderg sums of depedet radom varables, t s coveet to cosder the momet geeratg fucto (MGF). The MGF of a radom varable ca be represeted as follows (Arvats & Gregory, 00): M st ( s) = e pv ( t dt (6) V ) The well ow property of the MGF of a radom varables s that whe depedet varables are added, ther dstrbutos are covolved, but MGFs are multpled. Multplcato operato s easy to perform tha covoluto. O the other sde, oe has to obta the loss dstrbuto of from MGF usg verso tegrato. By sutably approxmatg the shape of the tegrad oe obtas a aalytcal approxmato of the probablty desty fucto. Ths

16 6 techque s ow as the method of steepest descets or saddle-pot method. The method also allows obtag aalytcal approxmatos to the probablty wthout havg to tegrate the desty fucto. The saddle-pot approxmato does ot mae ay pror assumptos about the shape of the loss dstrbuto. As was suggested by Arvats ad Gregory (00), saddle-pot approxmato method s a fast method obtag approxmate loss dstrbuto ad tal probabltes, ca be used for approxmato to derve expressos for loss dstrbutos that clude varable exposures ad default probabltes, ad allows corporatg correlato to the loss dstrbuto (p. 85). If we deote dvdual losses V,...,V be characterzed through ts cumulat geeratg fucto (Glasserma, 008): Each V, the dstrbuto of each loss ca ( exp( θ )) θ & ( =,..., ) Λ ( θ ) = log E (7) V s a radom fracto of a largest possble loss upo default of oblgor, all oblgors cosdered here are depedet. As a cosequece of depedece, the momet geeratg fucto ca be represeted as L E (8) = = ( θ ) = expθ YV = E[ exp( θyv )] = ( + p ( exp( Λ ( θ )) ) ) ad the cumulat geeratg fucto of L as (Glasserma, 008) ψ L ( θ ) = log[ L ( θ )] = log ( + p ( exp( Λ ( θ )) ) ) = ( log( + p ( exp( Λ ( θ )) ) )) (9) = = Aalyss of ths fucto shows that t s creasg, covex, ftely dfferetable ad taes zero whe = 0 θ. For a loss level x > 0, the saddle-pot s the root θ x of the equato ψ ' ( ) L θ x = x whch s uque. Cosder the cumulat geeratg fucto ψ L ( θ ). The dervatves of ( θ ) cumulats of L. The frst cumulat whe θ = 0 s the mea ψ L gve

17 7 ' L = = = ' ( 0) = p Λ ( 0) = p E[ V ] E[ L] ψ. (30) Whe losses are assumed to be costat, V = v, we have ( θ ) = v = ' ψ L p. (3) I other words, the dervatve of the cumulat geeratg fucto s the expected loss where orgal default probabltes are replaced wth p I geeral, the formula for dervatve s p = + p exp( v) ( exp() v ). ' ( θ ) ( θ ) ( θ ) ' ψ = p Λ = (3) whch ca be terpreted as the expected loss whe the orgal default probabltes are replaced p exp( Λ ( θ )) by p ( θ ) = ad the expected losses [ V ] + p ( exp( Λ ( θ )) ) ( ) ' E are replaced wth θ where Λ orgal expected loss [ V ] ' E cocdes wth Λ ( 0) because Λ s the cumulat geeratg fucto of V.Thus, each value of θ determes a modfed set of default probabltes ad a modfed loss gve default for each oblgor (Glasserma, 008, pp ). Due to complexty of the formulas provded, t s mportat to ether develop a fast algorthm or approxmato formulas. A saddle-pot approxmato ca be represeted the followg way (Glasserma, 008): P ( θ ) θ ψ " Φ ( θ ) ( L x ) ψ ( θ ) ψ " exp ( θ ) x x L x L x ( x L x ) > (33) ad the closely related Lugaa-Rce approxmato s P (34) λ( x) r( x) ( L > x) Φ( r( x) ) + ( r( x) )

18 8 where r( x) = ( θ ψ ( θ )) ad " λ x) θ ψ ( θ ) x x A modfcato of the prevous formula s P L x ( =. ( L > x) Φr( x) + log x r( x) L x λ. (35) r( x) Ths modfcato has the advatage that t always produces a value betwee 0 ad (Glasserma, 008, p. 456). To geerate the ucodtoal loss dstrbuto, as t was show before, oe has to compute a codtoal loss dstrbuto based o the assumpto of oblgor depedece ad the tegrate over the possble values of maret factors. Factor models are popular sce oblgors are codtoally depedet such models. Whe tegratg over the possble values of maret factors, fast tegrato procedures should be used; these procedures deped o the probablty dstrbuto of the chose maret factors ad the umber of maret factors. Glasserma (008) proposed ad algorthm for geerato of ucodtoal loss dstrbuto. The model specfes two sets of parameters correspodg to a hgh-default regme ad a low-default regme, wth depedet oblgors each regme. The the model uses a mxture of the two sets of parameters. I ths case, the uderlyg factor s the regme ad the ucodtoal loss dstrbuto may be computed as a mxture of the two codtoal loss dstrbutos (p. 457). Ths approach s qute useful whe aalyzg the credt portfolo cosstg of over 00 0 credts that ca perform dfferetly over the specfed tme perod. The credt portfolo ca be parttoed to the clusters (or bucets) of credts (as well as possble outlers) ad for each cluster low default regme ad hgh default regme ca be detfed. Glasserma ad Ruz-Mata (006) compared the computatoal effcecy of ordary Mote Carlo smulato wth methods that combe smulato for the factors wth the

19 9 techques used to geerate codtoal loss dstrbutos. They determed that umercal trasform verso ad saddle-pot approxmato volve some error the calculato of codtoal default probabltes. Because each replcato usg covoluto, trasform verso, or saddle-pot approxmato taes loger tha each replcato usg ordary smulato, these methods complete fewer replcatos a fxed amout of computg tme. The recursve covoluto method computes the full codtoal loss dstrbuto o each replcato whle trasform verso ad saddle-pot approxmato must practce be lmted to a smaller umber of loss thresholds. Usg the saddle-pot approxmato requres solvg for multple saddle-pot parameters o each replcato. The computato tme requred usg recursve covoluto grows qucly wth the umber of oblgors. As a cosequece of these, wth the total computg tme held fxed, ordary Mote Carlo ofte produces a smaller mea square error (Glasserma, 008, p. 458). The umber of factors plays a substatal role choosg whether Mote Carlo or aother method such as saddle-pot or recursve approach should be used. Whe the umber of factors s small, Mote Carlo smulatos ca be replaced by tegrato. For the moderate umber of dmesos, a quas-mote Carlo samplg ca be appled. O the other sde, oe ca use approxmatos where a sgle most mportat value of the factors s used (Glasserma, 004). Zheg (007) suggested approxmato of the codtoal loss dstrbuto usg a ormal dstrbuto by matchg two momets ad the computg the ucodtoal dstrbuto through umercal tegrato assumg small umber of factors. Specfcally Zheg (007) suggest computg mea ad varace depedg o maret factor Z: μ( Z) = p ( Z) ad σ Z = p ( Z) p ( Z) (36) = ( ) ( ) so that the 0 th ad frst order approxmato ca be represeted respectvely the followg way (Edgworth expaso): =

20 0 P P [ L x] x E Φ σ μ( Z ) ( Z ) x μ( z) x μ( z) [ L x Z] E Φ + H σ ( z) σ ( z) (37) (38) where H ( x) = (/ 6) σ ( z) γ ( z) = = 3 γ ( z)( x ) ( x) p ( z)( p ( z))( p ( z)) (39) The hgher order expaso ca be expressed usg Chebyshev-Hermte polyomals ad / momets of up to order +. The approxmato error was estmated as ( ) o. The ucodtoal loss dstrbuto the ca be obtaed usg tegrato. Glasserma ad L (005) proposed a two-step-mportace samplg techque that helps compute the upper tal of the loss dstrbuto. Such techques clude Mote-Carlo smulato methods combed wth adequate varace reducto approach, ad are applcable to a wde rage of applcatos. Importace samplg s a specal varace reducto techque ad s used to chage the dstrbutos of the relevat rs factors such a way that more realzatos of the loss radom varable are the upper tal. The each realzato s weghted by the lelhood rato to correct for the chage dstrbuto. Dembo et al. (004) provded a large devato approxmato of the tal dstrbuto of total facal losses o a portfolo cosstg of may postos. Quattatve aalyss of large losses s helpful structurg large portfolo so as to wthstad severe losses. Applcatos of ths approach clude the total default losses o a ba portfolo or the total clams agast a surer. A ey assumpto was that codtoal o a commo correlatg factor, posto looses are depedet. For large losses, facal dstress costs are more severe f the losses occur over

21 a relatvely short perod of tme. Sudde losses may cause extreme cash-flow stress, ad vestors may requre more favorable terms whe offerg ew les of facg over short tme perods, wth whch they may have a lmted opportuty to gather formato about the credt qualty ad log-term prospect of a dstressed facal sttutos. The results provded by the authors clude codtos uder whch a large-devatos estmate of the lelhood of a falurethreateg loss durg some sub-terval of tme durg a gve plag horzo ca be calculated from the lelhood of the same sze loss a certa fxed ey tme horzo. The codtoal dstrbuto of losses o each type of posto ca be estmated gve the large portfolo loss of cocer. The authors provded some aalytcal gudace o the depedece of large-loss probabltes o the structure of a portfolo wth a large umber of postos ad the most lely way that a large loss ca occur. Gve a large loss, the codtoal lelhood of loss o each type of posto ad codtoal dstrbuto of exposure the evet of loss were calculated. These codtoal calculatos ca be terpreted the asymptotc sese of Gbbs codtog prcple. Sdeus et al. (008) preseted the SPA framewor whch models are specfed by a two-layer process. The frst layer models the dyamcs of portfolo loss dstrbutos the absece of formato about default tmes. Ths bacgroud process ca be explctly calbrated to the full grd of margal loss dstrbutos as mpled by tal CDO trache values dexed o maturty, as well as to the prces of sutable optos. The authors gave suffcet codtos for cosstet dyamcs. The secod layer models the loss process tself as a Marov process codtoed o the path tae by the bacgroud process. The choce of loss process s ouque. Sdeus et al. (008) preseted a umber of choces, ad dscussed ther advatages ad dsadvatages. Several cocrete model examples were gve, ad valuato the ew

22 framewor was descrbed detal. Amog the specfc securtes for whch algorthms are preseted were CDO trache optos ad leveraged super-seor traches. I practce, oe ca estmate the mpled loss dstrbuto from observed maret data. For example, Kreel ad Partehemer (006) descrbed how to determe the mpled loss surface of a credt portfolo from CDO trache quotes. The approach ca be appled for prcg of CDO traches ad N th -to-default swaps ad for rs maagemet of CDO traches. It also ca serve as a tal dstrbuto for dyamc loss models. The calbrato ca be performed umercally by solvg a olear optmzato problem usg sequetal quadratc programmg method. Aalyss, cocluso ad recommedatos The most mportat ssue geeratg loss dstrbutos for credt portfolos s to fd a fast ad accurate algorthm whch ca be easly mplemeted. There s always trade-off betwee these characterstcs of the descrbed algorthms. Some of the aalyzed algorthms ca be mplemeted easly for solvg practcal problems; o the other sde, the dsadvatage of mplemetg them s a umber of assumptos that ca mae the model urealstc currg accurate decso mag. For a gve credt portfolo, t s advatageous cosderg varous loss dstrbutos models such as bucetg approach by Hull ad Whte (004), recursve method by Aderse et al. (003), Fourer aalytcal approach, ad saddle-pot approxmato ad use the oes that provde better soluto to the specfc problem ad the gve credt portfolo. Amog the algorthms descrbed above, the fastest algorthm s LHP, but the accuracy of ths algorthm depeds o the umber of oblgors the portfolo ad the credt spread dstrbuto. The more cocetrated the credt spreads are ad the more oblgors the credt

23 3 portfolo has, the more applcable LHP s. If we have a smaller umber of oblgors (say, less tha 00 oblgors) we ca use fte homogeeous approach. Fte homogeeous approach s slower tha LHP sce we eed to tegrate codtoal loss dstrbuto over the maret factor to obta ucodtoal loss dstrbuto. Ufortuately, practce, homogeety assumpto ad assumpto that the portfolo s large are ot the case ad LHP ca be rejected by practtoers seeg more accurate loss dstrbuto geerato. LHP as well as fte homogeeous portfolo methods ca t be used as a approxmato of heterogeeous portfolos a loss dstrbuto geerated usg a recursve method for a heterogeeous portfolo ca be qute dfferet from the loss dstrbuto geerated usg LHP whe the put parameter for the LHP s the weghted average spread of the heterogeeous portfolo. O the other extreme, whe more accurate geerato of loss dstrbuto s requred, the best method s Mote Carlo smulato whch s, at the same tme, the slowest. The Mote Carlo smulato s especally valuable cases whe the level of heterogeety s very hgh, other words whe the credt spread dstrbuto s very dverse. Mote Carlo smulatos (as well as other loss geerato methods) ca be made faster ad wll requre less memory f we cluster credt default spreads to homogeeous parttos of credt spreads. Gve a heterogeeous credt portfolo, the tas s to dstrbute to a group of clusters such a way that the objects wth each cluster are homogeeous ad hghly correlated whle the clusters themselves are dfferet. I fact, ths tas has two dffcultes frstly, we eed to detfy a optmal partto of the credt spreads to dfferet clusters; secodly the spread clusters should be explctly determed. There are a umber of algorthms that allows clusterg of credt spreads K-meas, Daa, Clara, fuzzy aalyss, self-orgazg maps, model-based clusterg. There are also

24 4 teral ad stablty valdato based o teral ad stablty measures. Tag the credt spreads data ad clusterg partto as a put, teral measures are used to assess the qualty of clusterg based o trsc formato. Stablty measures assess clusterg cosstecy by comparg t wth the clusters obtaed after each colum s removed. However, clusterg methods should be alged wth the goals pursued by researcher-practtoer sce clusterg methods aloe may ot gve approprate umber of clusters for accurate prcg of credt dervatves. For example, f a estmato of sum of squares wth groups usg -meas cluster aalyss ad valdato measures suggest 4 clusters of credt spreads, ths umber of clusters may ot eough to estmate CDO far spreads for each trache gve the partcular level of tolerace. Ths s because we stll may have hgh values for the wth-group sum of squares, low level of homogeety each cluster or, equvaletly, hgh level of heterogeety wth the clusters. Hgh level of heterogeety requres further crease umber of clusters to acheve covergece of estmated CDO far spreads to the real oe. The optmal umber of clusters depeds of the trache beg prced, credt spread dstrbuto of the credt portfolos, ad the recovery rates of the oblgors. I practce, the umber of clusters may be greater tha the oe suggested by, for example, the -meas cluster aalyss; however ths umber of clusters stll wll be several tmes less tha the umber of oblgors the credt portfolo thereby creasg the speed of loss dstrbuto geerato. The dsadvatage of usg cluster aalyss s that addtoal tme s requred to perform such aalyss. The movemets of credt spreads should be motored, ad the cluster aalyss mght be requred to be performed after each swtch regme chage. I rs maagemet of traches, loss dstrbutos for the credt portfolos eed to be extesvely recalculated for estmato of credt rs measures such as, for example, dosycratc

25 5 delta of a credt or a group of credts. Oe of the methods to speed up the buldg of loss dstrbuto s perturbato method suggested by O Kae (008, pp ). The perturbato method coupled wth parttog of credt spreads ca help further speed up the loss dstrbuto buldg ad allows more effcet estmato of sestvty of CDO trache prces to the stataeous movemet credt spreads of the cluster. Ths s because such approach s more cosstet wth what we usually observe the maret whe chages spreads occur a group of hghly correlated spreads rather tha oe partcular credt. Not all methods may be easly mplemeted stuatos whe the portfolo s heterogeeous. For example, the well ow algorthm proposed by Aderse et al. (003) oe has to estmate the value of commo dvsor of all potetal losses. The algorthm depeds o the statstcal characterstcs of the credt spreads (whch ca be dverse wth hgh level of heterogeety) ad some cases the value of the commo dvsor ca be very small substatally affectg the speed of the algorthm ad requrg more memory for estmato of the loss dstrbuto. A tolerace level of the commo dvsor of all potetal losses ca be chose to speed up the process, but ths case the resultg loss dstrbuto depeds o the chose tolerace level, whch may cur dscreteess of the loss dstrbuto shape ad eve may ot be accurate affectg credt rs maagemet ad credt dervatves prcg. Easess of mplemetato of specfc algorthms as well as accuracy depeds o the access of specalzed mathematcal lbrares. For example, to use a complex FFT algorthm for loss dstrbuto geerato, a aalyst would eed a access to a effectve FFT algorthm. Moreover, ths algorthm requres more tme for geerato of loss dstrbuto tha the recursve algorthms descrbed above ad ths oe of the ma reasos why FFT algorthms are ot wdely used practce. Whe estmatg a loss dstrbuto usg fte homogeeous portfolo, a

26 6 aalyst eeds to choose a effectve tegrato method as well as the tegrato rages to reach the requred tolerace. Geerato of loss dstrbuto should be alged wth other quattatve ad qualtatve aalyss of the oblgors ad formed decso should be made order to properly estmate the put parameters. For example, credt qualty of each credt ca chage over the specfed perod of tme, ad, therefore, t ca affect the loss dstrbuto that wll chage durg ths perod of tme due to chages credt qualty of the oblgors. The aalyzed methods assumed costat recovery rates. The value for recovery rates s ofte assumed to be equal to 40%, durg facal crss eve less 0 to 0%. Moreover, these values are ofte assumed to be the same for all the oblgors wth dfferet ratgs ad credt spread term structure. Estmato of recovery rates for each oblgor based o facal statemet aalyss, for example, estmato of recovery rates for a portfolo cosstg of 40 oblgors s tme cosumg; o the other sde, the geerated loss dstrbuto based o the carefully estmated recovery rates would be more realstc. Sce recovery rates are chagg over the partcular tme perod, the loss dstrbuto geerato methods ca be mproved by assumg that recovery rates for each oblgor (ad, therefore, loss gve default) are radom varables. The probabltes of default ad recovery rates are egatvely correlated ad they ca be exogeously modeled as stochastc processes whch the ca be corporated to exstg loss dstrbuto geerato models. The descrbed methods do t cosder credt spread dstrbuto of the credt portfolo. Loss dstrbuto of the portfolo or CDO trache deped o the credt spread dstrbutos observed the maret. Aalyss of credt spread dstrbutos ca help detfy whch credt spreads are most lely affect specfc CDO trache, ad whch credt spreads would t. For

27 7 example, for a gve teor, say 3 years, a credt portfolo may have credts wth low credt spreads cocetratg aroud the value gvg tme to default whch are more tha 0 years. Such credt spreads do t affect loss dstrbuto of the cosdered CDO trache, ad, therefore ca be excluded from cosderato. If a umber of such credt spreads s very hgh, the by excludg these credts from cosderato we ca substatally decrease computatoal tme of the loss dstrbuto geerato. I other words, loss dstrbuto geerato ca be faster ad substatally mproved, f we perform prelmary statstcal aalyss of credt spread dstrbuto, recovery rates ad default probabltes of the oblgors.

28 8 Refereces Aderse, L. J., Sdeus, J. & Basu, S. (003, November). All your hedges oe baset. Rs, Arvats, A. & Gregory, J. (999). A credt rs toolbox. I D. Shmo (ed.), Credt rs. Models ad maagemet. pp Lodo: Rs boos. Arvats, A. & Gregory, J. (00). Credt: The complete gude to prcg, hedgg ad rs maagemet. Lodo: Rs Boos. Dembo, A., Deuschel, J., & Duffe, D. (004). Large portfolo losses. Face & Stochastcs, 8(), 3-6. Glasserma, P. & L, J. (003). Importace samplg for a mxed Posso model of portfolo credt rs. I: S. Chc, P.J. Sachez, D. Ferr, ad D.J. Morrse (Eds.), Proceedgs of the 003 Wter Smulato Coferece, pp Glasserma, P. & Ruz-Mata, J. (006). A comparso of approxmato techques for portfolo credt rs. Joural of Credt Rs, Glasserma, P. (004). Tal approxmatos for portfolo credt rs. Joural of Dervatves, 4-4. Gordy, M. (003, July). A rs-factor model foudato for ratgs-based ba captal rules. Joural of Facal Itermedato, (3), Gordy, M. (004). Graularty adjustmet portfolo rs measuremet. I G. Szegö (ed.), Rs Measures for the st Cetury. pp Joh Wley & Sos. Gordy,M.B. (00). Saddlepot approxmato of CredtRs+. Joural of Bag ad Face 6(7), Gregory, J., & Lauret, J. (003, Jue). I wll survve. Rs, 6(6), Gregory, J., & Lauret, J. (004, October). I the core of correlato. Rs 7(0), 87-9 Grude, P.(007). Computatoal aspects of tegrated maret ad credt portfolo models. OR Spectrum, 9(), Hull, J. & Whte, A. (004). Valuato of a CDO ad a th to default CDS wthout Mote Carlo smulatos. Joural of Dervatves, (), 8-3 Kalemaova, A., Schmd, B. & Werer, R. (007, September). The ormal verse Gaussa dstrbuto for sythetc CDO Prcg. Joural of Dervatves, 4(3), pp Kreel, M. & Partehemer, J. (006). The mpled loss surface of CDOs. Retreved Jauary 6, 009, from

29 9 Lauret, J., & Gregory, J. (005, Summer). Baset default swaps, CDOs ad factor copulas. Joural of Rs, 7(4), -0. Lauret, J.-P. (004). Fast aalytc techques for prcg sythetc CDOs. Presetato at Credt Rs Summt Europe. October 3, 004. Mart, R. (006, December). The saddlepot approxmato ad portfolo optoaltes. Rs, Mero, S. & Nyfeler, M. (00, August). Calculatg portfolo loss. Rs 5(8), 8-86 O Kae, D. (008). Modelg sgle-ame ad mult-ame credt dervatves. New Yor, NY: Joh Wley & Sos Pyht, M. (004, March). Mult-factor adjustmet. Rs 7(3) Reβ, O. (003). Mathematcal methods for the effcet assessmet of maret ad credt rs. PhD Thess. Departmet of Mathematcs, Uversty of Kaserslauter. Schloegl, L., & O'Kae, D. (005, October). A ote o the large homogeeous portfolo approxmato wth the Studet-t copula. Face & Stochastcs, 9(4), Schöbucher, P.J. (00). Tae to the lmt: Smple ad ot so smple loa loss dstrbutos. Worg Paper, Departmet of Statstcs, Bo Uversty. Sdeus, J., Pterbarg, V., & Aderse, L. (008, March). A ew framewor for dyamc credt portfolo loss modelg. Iteratoal Joural of Theoretcal & Appled Face, (), Vasce, O. (987). Probablty of loss o loa portfolo. Worg paper, KMV Corporato. Vasce, O. (00, December). Probablty of loss o a loa portfolo. Rs, Zheg, H. (007). Portfolo dstrbuto modelg ad computato. Worshop o Fast Facal Algorthms. Taaa Busess School & Imperal College. Retreved Jauary 0, 009, from

An Effectiveness of Integrated Portfolio in Bancassurance

An Effectiveness of Integrated Portfolio in Bancassurance A Effectveess of Itegrated Portfolo Bacassurace Taea Karya Research Ceter for Facal Egeerg Isttute of Ecoomc Research Kyoto versty Sayouu Kyoto 606-850 Japa arya@eryoto-uacp Itroducto As s well ow the

More information

T = 1/freq, T = 2/freq, T = i/freq, T = n (number of cash flows = freq n) are :

T = 1/freq, T = 2/freq, T = i/freq, T = n (number of cash flows = freq n) are : Bullets bods Let s descrbe frst a fxed rate bod wthout amortzg a more geeral way : Let s ote : C the aual fxed rate t s a percetage N the otoal freq ( 2 4 ) the umber of coupo per year R the redempto of

More information

Average Price Ratios

Average Price Ratios Average Prce Ratos Morgstar Methodology Paper August 3, 2005 2005 Morgstar, Ic. All rghts reserved. The formato ths documet s the property of Morgstar, Ic. Reproducto or trascrpto by ay meas, whole or

More information

ANOVA Notes Page 1. Analysis of Variance for a One-Way Classification of Data

ANOVA Notes Page 1. Analysis of Variance for a One-Way Classification of Data ANOVA Notes Page Aalss of Varace for a Oe-Wa Classfcato of Data Cosder a sgle factor or treatmet doe at levels (e, there are,, 3, dfferet varatos o the prescrbed treatmet) Wth a gve treatmet level there

More information

STATISTICAL PROPERTIES OF LEAST SQUARES ESTIMATORS. x, where. = y - ˆ " 1

STATISTICAL PROPERTIES OF LEAST SQUARES ESTIMATORS. x, where. = y - ˆ  1 STATISTICAL PROPERTIES OF LEAST SQUARES ESTIMATORS Recall Assumpto E(Y x) η 0 + η x (lear codtoal mea fucto) Data (x, y ), (x 2, y 2 ),, (x, y ) Least squares estmator ˆ E (Y x) ˆ " 0 + ˆ " x, where ˆ

More information

Abraham Zaks. Technion I.I.T. Haifa ISRAEL. and. University of Haifa, Haifa ISRAEL. Abstract

Abraham Zaks. Technion I.I.T. Haifa ISRAEL. and. University of Haifa, Haifa ISRAEL. Abstract Preset Value of Autes Uder Radom Rates of Iterest By Abraham Zas Techo I.I.T. Hafa ISRAEL ad Uversty of Hafa, Hafa ISRAEL Abstract Some attempts were made to evaluate the future value (FV) of the expected

More information

1. The Time Value of Money

1. The Time Value of Money Corporate Face [00-0345]. The Tme Value of Moey. Compoudg ad Dscoutg Captalzato (compoudg, fdg future values) s a process of movg a value forward tme. It yelds the future value gve the relevat compoudg

More information

6.7 Network analysis. 6.7.1 Introduction. References - Network analysis. Topological analysis

6.7 Network analysis. 6.7.1 Introduction. References - Network analysis. Topological analysis 6.7 Network aalyss Le data that explctly store topologcal formato are called etwork data. Besdes spatal operatos, several methods of spatal aalyss are applcable to etwork data. Fgure: Network data Refereces

More information

SHAPIRO-WILK TEST FOR NORMALITY WITH KNOWN MEAN

SHAPIRO-WILK TEST FOR NORMALITY WITH KNOWN MEAN SHAPIRO-WILK TEST FOR NORMALITY WITH KNOWN MEAN Wojcech Zelńsk Departmet of Ecoometrcs ad Statstcs Warsaw Uversty of Lfe Sceces Nowoursyowska 66, -787 Warszawa e-mal: wojtekzelsk@statystykafo Zofa Hausz,

More information

Credibility Premium Calculation in Motor Third-Party Liability Insurance

Credibility Premium Calculation in Motor Third-Party Liability Insurance Advaces Mathematcal ad Computatoal Methods Credblty remum Calculato Motor Thrd-arty Lablty Isurace BOHA LIA, JAA KUBAOVÁ epartmet of Mathematcs ad Quattatve Methods Uversty of ardubce Studetská 95, 53

More information

APPENDIX III THE ENVELOPE PROPERTY

APPENDIX III THE ENVELOPE PROPERTY Apped III APPENDIX III THE ENVELOPE PROPERTY Optmzato mposes a very strog structure o the problem cosdered Ths s the reaso why eoclasscal ecoomcs whch assumes optmzg behavour has bee the most successful

More information

Reinsurance and the distribution of term insurance claims

Reinsurance and the distribution of term insurance claims Resurace ad the dstrbuto of term surace clams By Rchard Bruyel FIAA, FNZSA Preseted to the NZ Socety of Actuares Coferece Queestow - November 006 1 1 Itroducto Ths paper vestgates the effect of resurace

More information

The Gompertz-Makeham distribution. Fredrik Norström. Supervisor: Yuri Belyaev

The Gompertz-Makeham distribution. Fredrik Norström. Supervisor: Yuri Belyaev The Gompertz-Makeham dstrbuto by Fredrk Norström Master s thess Mathematcal Statstcs, Umeå Uversty, 997 Supervsor: Yur Belyaev Abstract Ths work s about the Gompertz-Makeham dstrbuto. The dstrbuto has

More information

IDENTIFICATION OF THE DYNAMICS OF THE GOOGLE S RANKING ALGORITHM. A. Khaki Sedigh, Mehdi Roudaki

IDENTIFICATION OF THE DYNAMICS OF THE GOOGLE S RANKING ALGORITHM. A. Khaki Sedigh, Mehdi Roudaki IDENIFICAION OF HE DYNAMICS OF HE GOOGLE S RANKING ALGORIHM A. Khak Sedgh, Mehd Roudak Cotrol Dvso, Departmet of Electrcal Egeerg, K.N.oos Uversty of echology P. O. Box: 16315-1355, ehra, Ira sedgh@eetd.ktu.ac.r,

More information

Statistical Pattern Recognition (CE-725) Department of Computer Engineering Sharif University of Technology

Statistical Pattern Recognition (CE-725) Department of Computer Engineering Sharif University of Technology I The Name of God, The Compassoate, The ercful Name: Problems' eys Studet ID#:. Statstcal Patter Recogto (CE-725) Departmet of Computer Egeerg Sharf Uversty of Techology Fal Exam Soluto - Sprg 202 (50

More information

Numerical Methods with MS Excel

Numerical Methods with MS Excel TMME, vol4, o.1, p.84 Numercal Methods wth MS Excel M. El-Gebely & B. Yushau 1 Departmet of Mathematcal Sceces Kg Fahd Uversty of Petroleum & Merals. Dhahra, Saud Araba. Abstract: I ths ote we show how

More information

Report 52 Fixed Maturity EUR Industrial Bond Funds

Report 52 Fixed Maturity EUR Industrial Bond Funds Rep52, Computed & Prted: 17/06/2015 11:53 Report 52 Fxed Maturty EUR Idustral Bod Fuds From Dec 2008 to Dec 2014 31/12/2008 31 December 1999 31/12/2014 Bechmark Noe Defto of the frm ad geeral formato:

More information

ADAPTATION OF SHAPIRO-WILK TEST TO THE CASE OF KNOWN MEAN

ADAPTATION OF SHAPIRO-WILK TEST TO THE CASE OF KNOWN MEAN Colloquum Bometrcum 4 ADAPTATION OF SHAPIRO-WILK TEST TO THE CASE OF KNOWN MEAN Zofa Hausz, Joaa Tarasńska Departmet of Appled Mathematcs ad Computer Scece Uversty of Lfe Sceces Lubl Akademcka 3, -95 Lubl

More information

Maintenance Scheduling of Distribution System with Optimal Economy and Reliability

Maintenance Scheduling of Distribution System with Optimal Economy and Reliability Egeerg, 203, 5, 4-8 http://dx.do.org/0.4236/eg.203.59b003 Publshed Ole September 203 (http://www.scrp.org/joural/eg) Mateace Schedulg of Dstrbuto System wth Optmal Ecoomy ad Relablty Syua Hog, Hafeg L,

More information

RUSSIAN ROULETTE AND PARTICLE SPLITTING

RUSSIAN ROULETTE AND PARTICLE SPLITTING RUSSAN ROULETTE AND PARTCLE SPLTTNG M. Ragheb 3/7/203 NTRODUCTON To stuatos are ecoutered partcle trasport smulatos:. a multplyg medum, a partcle such as a eutro a cosmc ray partcle or a photo may geerate

More information

A New Bayesian Network Method for Computing Bottom Event's Structural Importance Degree using Jointree

A New Bayesian Network Method for Computing Bottom Event's Structural Importance Degree using Jointree , pp.277-288 http://dx.do.org/10.14257/juesst.2015.8.1.25 A New Bayesa Network Method for Computg Bottom Evet's Structural Importace Degree usg Jotree Wag Yao ad Su Q School of Aeroautcs, Northwester Polytechcal

More information

Performance Attribution. Methodology Overview

Performance Attribution. Methodology Overview erformace Attrbuto Methodology Overvew Faba SUAREZ March 2004 erformace Attrbuto Methodology 1.1 Itroducto erformace Attrbuto s a set of techques that performace aalysts use to expla why a portfolo's performace

More information

CHAPTER 2. Time Value of Money 6-1

CHAPTER 2. Time Value of Money 6-1 CHAPTER 2 Tme Value of Moey 6- Tme Value of Moey (TVM) Tme Les Future value & Preset value Rates of retur Autes & Perpetutes Ueve cash Flow Streams Amortzato 6-2 Tme les 0 2 3 % CF 0 CF CF 2 CF 3 Show

More information

ECONOMIC CHOICE OF OPTIMUM FEEDER CABLE CONSIDERING RISK ANALYSIS. University of Brasilia (UnB) and The Brazilian Regulatory Agency (ANEEL), Brazil

ECONOMIC CHOICE OF OPTIMUM FEEDER CABLE CONSIDERING RISK ANALYSIS. University of Brasilia (UnB) and The Brazilian Regulatory Agency (ANEEL), Brazil ECONOMIC CHOICE OF OPTIMUM FEEDER CABE CONSIDERING RISK ANAYSIS I Camargo, F Fgueredo, M De Olvera Uversty of Brasla (UB) ad The Brazla Regulatory Agecy (ANEE), Brazl The choce of the approprate cable

More information

The Analysis of Development of Insurance Contract Premiums of General Liability Insurance in the Business Insurance Risk

The Analysis of Development of Insurance Contract Premiums of General Liability Insurance in the Business Insurance Risk The Aalyss of Developmet of Isurace Cotract Premums of Geeral Lablty Isurace the Busess Isurace Rsk the Frame of the Czech Isurace Market 1998 011 Scetfc Coferece Jue, 10. - 14. 013 Pavla Kubová Departmet

More information

Online Appendix: Measured Aggregate Gains from International Trade

Online Appendix: Measured Aggregate Gains from International Trade Ole Appedx: Measured Aggregate Gas from Iteratoal Trade Arel Burste UCLA ad NBER Javer Cravo Uversty of Mchga March 3, 2014 I ths ole appedx we derve addtoal results dscussed the paper. I the frst secto,

More information

Models for Selecting an ERP System with Intuitionistic Trapezoidal Fuzzy Information

Models for Selecting an ERP System with Intuitionistic Trapezoidal Fuzzy Information JOURNAL OF SOFWARE, VOL 5, NO 3, MARCH 00 75 Models for Selectg a ERP System wth Itutostc rapezodal Fuzzy Iformato Guwu We, Ru L Departmet of Ecoomcs ad Maagemet, Chogqg Uversty of Arts ad Sceces, Yogchua,

More information

Chapter 3. AMORTIZATION OF LOAN. SINKING FUNDS R =

Chapter 3. AMORTIZATION OF LOAN. SINKING FUNDS R = Chapter 3. AMORTIZATION OF LOAN. SINKING FUNDS Objectves of the Topc: Beg able to formalse ad solve practcal ad mathematcal problems, whch the subjects of loa amortsato ad maagemet of cumulatve fuds are

More information

Optimal multi-degree reduction of Bézier curves with constraints of endpoints continuity

Optimal multi-degree reduction of Bézier curves with constraints of endpoints continuity Computer Aded Geometrc Desg 19 (2002 365 377 wwwelsevercom/locate/comad Optmal mult-degree reducto of Bézer curves wth costrats of edpots cotuty Guo-Dog Che, Guo-J Wag State Key Laboratory of CAD&CG, Isttute

More information

Preprocess a planar map S. Given a query point p, report the face of S containing p. Goal: O(n)-size data structure that enables O(log n) query time.

Preprocess a planar map S. Given a query point p, report the face of S containing p. Goal: O(n)-size data structure that enables O(log n) query time. Computatoal Geometry Chapter 6 Pot Locato 1 Problem Defto Preprocess a plaar map S. Gve a query pot p, report the face of S cotag p. S Goal: O()-sze data structure that eables O(log ) query tme. C p E

More information

n. We know that the sum of squares of p independent standard normal variables has a chi square distribution with p degrees of freedom.

n. We know that the sum of squares of p independent standard normal variables has a chi square distribution with p degrees of freedom. UMEÅ UNIVERSITET Matematsk-statstska sttutoe Multvarat dataaalys för tekologer MSTB0 PA TENTAMEN 004-0-9 LÖSNINGSFÖRSLAG TILL TENTAMEN I MATEMATISK STATISTIK Multvarat dataaalys för tekologer B, 5 poäg.

More information

Settlement Prediction by Spatial-temporal Random Process

Settlement Prediction by Spatial-temporal Random Process Safety, Relablty ad Rs of Structures, Ifrastructures ad Egeerg Systems Furuta, Fragopol & Shozua (eds Taylor & Fracs Group, Lodo, ISBN 978---77- Settlemet Predcto by Spatal-temporal Radom Process P. Rugbaapha

More information

Applications of Support Vector Machine Based on Boolean Kernel to Spam Filtering

Applications of Support Vector Machine Based on Boolean Kernel to Spam Filtering Moder Appled Scece October, 2009 Applcatos of Support Vector Mache Based o Boolea Kerel to Spam Flterg Shugag Lu & Keb Cu School of Computer scece ad techology, North Cha Electrc Power Uversty Hebe 071003,

More information

USEFULNESS OF BOOTSTRAPPING IN PORTFOLIO MANAGEMENT

USEFULNESS OF BOOTSTRAPPING IN PORTFOLIO MANAGEMENT USEFULNESS OF BOOTSTRAPPING IN PORTFOLIO MANAGEMENT Radovaov Bors Faculty of Ecoomcs Subotca Segedsk put 9-11 Subotca 24000 E-mal: radovaovb@ef.us.ac.rs Marckć Aleksadra Faculty of Ecoomcs Subotca Segedsk

More information

We present a new approach to pricing American-style derivatives that is applicable to any Markovian setting

We present a new approach to pricing American-style derivatives that is applicable to any Markovian setting MANAGEMENT SCIENCE Vol. 52, No., Jauary 26, pp. 95 ss 25-99 ess 526-55 6 52 95 forms do.287/msc.5.447 26 INFORMS Prcg Amerca-Style Dervatves wth Europea Call Optos Scott B. Laprse BAE Systems, Advaced

More information

Fractal-Structured Karatsuba`s Algorithm for Binary Field Multiplication: FK

Fractal-Structured Karatsuba`s Algorithm for Binary Field Multiplication: FK Fractal-Structured Karatsuba`s Algorthm for Bary Feld Multplcato: FK *The authors are worg at the Isttute of Mathematcs The Academy of Sceces of DPR Korea. **Address : U Jog dstrct Kwahadog Number Pyogyag

More information

Simple Linear Regression

Simple Linear Regression Smple Lear Regresso Regresso equato a equato that descrbes the average relatoshp betwee a respose (depedet) ad a eplaator (depedet) varable. 6 8 Slope-tercept equato for a le m b (,6) slope. (,) 6 6 8

More information

The Digital Signature Scheme MQQ-SIG

The Digital Signature Scheme MQQ-SIG The Dgtal Sgature Scheme MQQ-SIG Itellectual Property Statemet ad Techcal Descrpto Frst publshed: 10 October 2010, Last update: 20 December 2010 Dalo Glgorosk 1 ad Rue Stesmo Ødegård 2 ad Rue Erled Jese

More information

Integrating Production Scheduling and Maintenance: Practical Implications

Integrating Production Scheduling and Maintenance: Practical Implications Proceedgs of the 2012 Iteratoal Coferece o Idustral Egeerg ad Operatos Maagemet Istabul, Turkey, uly 3 6, 2012 Itegratg Producto Schedulg ad Mateace: Practcal Implcatos Lath A. Hadd ad Umar M. Al-Turk

More information

RESEARCH ON PERFORMANCE MODELING OF TRANSACTIONAL CLOUD APPLICATIONS

RESEARCH ON PERFORMANCE MODELING OF TRANSACTIONAL CLOUD APPLICATIONS Joural of Theoretcal ad Appled Iformato Techology 3 st October 22. Vol. 44 No.2 25-22 JATIT & LLS. All rghts reserved. ISSN: 992-8645 www.jatt.org E-ISSN: 87-395 RESEARCH ON PERFORMANCE MODELING OF TRANSACTIONAL

More information

ANNEX 77 FINANCE MANAGEMENT. (Working material) Chief Actuary Prof. Gaida Pettere BTA INSURANCE COMPANY SE

ANNEX 77 FINANCE MANAGEMENT. (Working material) Chief Actuary Prof. Gaida Pettere BTA INSURANCE COMPANY SE ANNEX 77 FINANCE MANAGEMENT (Workg materal) Chef Actuary Prof. Gada Pettere BTA INSURANCE COMPANY SE 1 FUNDAMENTALS of INVESTMENT I THEORY OF INTEREST RATES 1.1 ACCUMULATION Iterest may be regarded as

More information

Optimal replacement and overhaul decisions with imperfect maintenance and warranty contracts

Optimal replacement and overhaul decisions with imperfect maintenance and warranty contracts Optmal replacemet ad overhaul decsos wth mperfect mateace ad warraty cotracts R. Pascual Departmet of Mechacal Egeerg, Uversdad de Chle, Caslla 2777, Satago, Chle Phoe: +56-2-6784591 Fax:+56-2-689657 rpascual@g.uchle.cl

More information

The Time Value of Money

The Time Value of Money The Tme Value of Moey 1 Iversemet Optos Year: 1624 Property Traded: Mahatta Islad Prce : $24.00, FV of $24 @ 6%: FV = $24 (1+0.06) 388 = $158.08 bllo Opto 1 0 1 2 3 4 5 t ($519.37) 0 0 0 0 $1,000 Opto

More information

On Error Detection with Block Codes

On Error Detection with Block Codes BULGARIAN ACADEMY OF SCIENCES CYBERNETICS AND INFORMATION TECHNOLOGIES Volume 9, No 3 Sofa 2009 O Error Detecto wth Block Codes Rostza Doduekova Chalmers Uversty of Techology ad the Uversty of Gotheburg,

More information

Proceedings of the 2010 Winter Simulation Conference B. Johansson, S. Jain, J. Montoya-Torres, J. Hugan, and E. Yücesan, eds.

Proceedings of the 2010 Winter Simulation Conference B. Johansson, S. Jain, J. Montoya-Torres, J. Hugan, and E. Yücesan, eds. Proceedgs of the 21 Wter Smulato Coferece B. Johasso, S. Ja, J. Motoya-Torres, J. Huga, ad E. Yücesa, eds. EMPIRICAL METHODS OR TWO-ECHELON INVENTORY MANAGEMENT WITH SERVICE LEVEL CONSTRAINTS BASED ON

More information

The analysis of annuities relies on the formula for geometric sums: r k = rn+1 1 r 1. (2.1) k=0

The analysis of annuities relies on the formula for geometric sums: r k = rn+1 1 r 1. (2.1) k=0 Chapter 2 Autes ad loas A auty s a sequece of paymets wth fxed frequecy. The term auty orgally referred to aual paymets (hece the ame), but t s ow also used for paymets wth ay frequecy. Autes appear may

More information

Report 19 Euroland Corporate Bonds

Report 19 Euroland Corporate Bonds Rep19, Computed & Prted: 17/06/2015 11:38 Report 19 Eurolad Corporate Bods From Dec 1999 to Dec 2014 31/12/1999 31 December 1999 31/12/2014 Bechmark 100% IBOXX Euro Corp All Mats. TR Defto of the frm ad

More information

Capacitated Production Planning and Inventory Control when Demand is Unpredictable for Most Items: The No B/C Strategy

Capacitated Production Planning and Inventory Control when Demand is Unpredictable for Most Items: The No B/C Strategy SCHOOL OF OPERATIONS RESEARCH AND INDUSTRIAL ENGINEERING COLLEGE OF ENGINEERING CORNELL UNIVERSITY ITHACA, NY 4853-380 TECHNICAL REPORT Jue 200 Capactated Producto Plag ad Ivetory Cotrol whe Demad s Upredctable

More information

An IG-RS-SVM classifier for analyzing reviews of E-commerce product

An IG-RS-SVM classifier for analyzing reviews of E-commerce product Iteratoal Coferece o Iformato Techology ad Maagemet Iovato (ICITMI 205) A IG-RS-SVM classfer for aalyzg revews of E-commerce product Jaju Ye a, Hua Re b ad Hagxa Zhou c * College of Iformato Egeerg, Cha

More information

DETERMINISTIC AND STOCHASTIC MODELLING OF TECHNICAL RESERVES IN SHORT-TERM INSURANCE CONTRACTS

DETERMINISTIC AND STOCHASTIC MODELLING OF TECHNICAL RESERVES IN SHORT-TERM INSURANCE CONTRACTS DETERMINISTI AND STOHASTI MODELLING OF TEHNIAL RESERVES IN SHORT-TERM INSURANE ONTRATS Patrck G O Weke School of Mathematcs, Uversty of Narob, Keya Emal: pweke@uobacke ABSTART lams reservg for geeral surace

More information

Commercial Pension Insurance Program Design and Estimated of Tax Incentives---- Based on Analysis of Enterprise Annuity Tax Incentives

Commercial Pension Insurance Program Design and Estimated of Tax Incentives---- Based on Analysis of Enterprise Annuity Tax Incentives Iteratoal Joural of Busess ad Socal Scece Vol 5, No ; October 204 Commercal Peso Isurace Program Desg ad Estmated of Tax Icetves---- Based o Aalyss of Eterprse Auty Tax Icetves Huag Xue, Lu Yatg School

More information

Forecasting Trend and Stock Price with Adaptive Extended Kalman Filter Data Fusion

Forecasting Trend and Stock Price with Adaptive Extended Kalman Filter Data Fusion 2011 Iteratoal Coferece o Ecoomcs ad Face Research IPEDR vol.4 (2011 (2011 IACSIT Press, Sgapore Forecastg Tred ad Stoc Prce wth Adaptve Exteded alma Flter Data Fuso Betollah Abar Moghaddam Faculty of

More information

Dynamic Two-phase Truncated Rayleigh Model for Release Date Prediction of Software

Dynamic Two-phase Truncated Rayleigh Model for Release Date Prediction of Software J. Software Egeerg & Applcatos 3 63-69 do:.436/jsea..367 Publshed Ole Jue (http://www.scrp.org/joural/jsea) Dyamc Two-phase Trucated Raylegh Model for Release Date Predcto of Software Lafe Qa Qgchua Yao

More information

AN ALGORITHM ABOUT PARTNER SELECTION PROBLEM ON CLOUD SERVICE PROVIDER BASED ON GENETIC

AN ALGORITHM ABOUT PARTNER SELECTION PROBLEM ON CLOUD SERVICE PROVIDER BASED ON GENETIC Joural of Theoretcal ad Appled Iformato Techology 0 th Aprl 204. Vol. 62 No. 2005-204 JATIT & LLS. All rghts reserved. ISSN: 992-8645 www.jatt.org E-ISSN: 87-395 AN ALGORITHM ABOUT PARTNER SELECTION PROBLEM

More information

How To Make A Supply Chain System Work

How To Make A Supply Chain System Work Iteratoal Joural of Iformato Techology ad Kowledge Maagemet July-December 200, Volume 2, No. 2, pp. 3-35 LATERAL TRANSHIPMENT-A TECHNIQUE FOR INVENTORY CONTROL IN MULTI RETAILER SUPPLY CHAIN SYSTEM Dharamvr

More information

of the relationship between time and the value of money.

of the relationship between time and the value of money. TIME AND THE VALUE OF MONEY Most agrbusess maagers are famlar wth the terms compoudg, dscoutg, auty, ad captalzato. That s, most agrbusess maagers have a tutve uderstadg that each term mples some relatoshp

More information

ANALYTICAL MODEL FOR TCP FILE TRANSFERS OVER UMTS. Janne Peisa Ericsson Research 02420 Jorvas, Finland. Michael Meyer Ericsson Research, Germany

ANALYTICAL MODEL FOR TCP FILE TRANSFERS OVER UMTS. Janne Peisa Ericsson Research 02420 Jorvas, Finland. Michael Meyer Ericsson Research, Germany ANALYTICAL MODEL FOR TCP FILE TRANSFERS OVER UMTS Jae Pesa Erco Research 4 Jorvas, Flad Mchael Meyer Erco Research, Germay Abstract Ths paper proposes a farly complex model to aalyze the performace of

More information

EMERGING MARKETS: STOCK MARKET INVESTING WITH POLITICAL RISK. Ephraim Clark and Radu Tunaru

EMERGING MARKETS: STOCK MARKET INVESTING WITH POLITICAL RISK. Ephraim Clark and Radu Tunaru EMERGING MARKETS: STOCK MARKET INVESTING WITH POLITICAL RISK By Ephram Clark ad Radu Tuaru Correspodace to: Ephram Clark Mddlesex Uversty Busess School The Burroughs Lodo NW4 4BT UK Tel: 44-(0)08-36-530

More information

DECISION MAKING WITH THE OWA OPERATOR IN SPORT MANAGEMENT

DECISION MAKING WITH THE OWA OPERATOR IN SPORT MANAGEMENT ESTYLF08, Cuecas Meras (Meres - Lagreo), 7-9 de Septembre de 2008 DECISION MAKING WITH THE OWA OPERATOR IN SPORT MANAGEMENT José M. Mergó Aa M. Gl-Lafuete Departmet of Busess Admstrato, Uversty of Barceloa

More information

Chapter 3 0.06 = 3000 ( 1.015 ( 1 ) Present Value of an Annuity. Section 4 Present Value of an Annuity; Amortization

Chapter 3 0.06 = 3000 ( 1.015 ( 1 ) Present Value of an Annuity. Section 4 Present Value of an Annuity; Amortization Chapter 3 Mathematcs of Face Secto 4 Preset Value of a Auty; Amortzato Preset Value of a Auty I ths secto, we wll address the problem of determg the amout that should be deposted to a accout ow at a gve

More information

RQM: A new rate-based active queue management algorithm

RQM: A new rate-based active queue management algorithm : A ew rate-based actve queue maagemet algorthm Jeff Edmods, Suprakash Datta, Patrck Dymod, Kashf Al Computer Scece ad Egeerg Departmet, York Uversty, Toroto, Caada Abstract I ths paper, we propose a ew

More information

Three Dimensional Interpolation of Video Signals

Three Dimensional Interpolation of Video Signals Three Dmesoal Iterpolato of Vdeo Sgals Elham Shahfard March 0 th 006 Outle A Bref reve of prevous tals Dgtal Iterpolato Bascs Upsamplg D Flter Desg Issues Ifte Impulse Respose Fte Impulse Respose Desged

More information

Location Analysis Regarding Disaster Management Bases via GIS Case study: Tehran Municipality (No.6)

Location Analysis Regarding Disaster Management Bases via GIS Case study: Tehran Municipality (No.6) Urba - Regoal Studes ad Research Joural 3 rd Year No.10 - Autum 2011 Locato Aalyss Regardg Dsaster Maagemet Bases va GIS Case study: Tehra Mucpalty (No.6) M. Shoja Aragh. S. Tavallae. P. Zaea Receved:

More information

The simple linear Regression Model

The simple linear Regression Model The smple lear Regresso Model Correlato coeffcet s o-parametrc ad just dcates that two varables are assocated wth oe aother, but t does ot gve a deas of the kd of relatoshp. Regresso models help vestgatg

More information

ISyE 512 Chapter 7. Control Charts for Attributes. Instructor: Prof. Kaibo Liu. Department of Industrial and Systems Engineering UW-Madison

ISyE 512 Chapter 7. Control Charts for Attributes. Instructor: Prof. Kaibo Liu. Department of Industrial and Systems Engineering UW-Madison ISyE 512 Chapter 7 Cotrol Charts for Attrbutes Istructor: Prof. Kabo Lu Departmet of Idustral ad Systems Egeerg UW-Madso Emal: klu8@wsc.edu Offce: Room 3017 (Mechacal Egeerg Buldg) 1 Lst of Topcs Chapter

More information

Network dimensioning for elastic traffic based on flow-level QoS

Network dimensioning for elastic traffic based on flow-level QoS Network dmesog for elastc traffc based o flow-level QoS 1(10) Network dmesog for elastc traffc based o flow-level QoS Pas Lassla ad Jorma Vrtamo Networkg Laboratory Helsk Uversty of Techology Itroducto

More information

Green Master based on MapReduce Cluster

Green Master based on MapReduce Cluster Gree Master based o MapReduce Cluster Mg-Zh Wu, Yu-Chag L, We-Tsog Lee, Yu-Su L, Fog-Hao Lu Dept of Electrcal Egeerg Tamkag Uversty, Tawa, ROC Dept of Electrcal Egeerg Tamkag Uversty, Tawa, ROC Dept of

More information

Analysis of Multi-product Break-even with Uncertain Information*

Analysis of Multi-product Break-even with Uncertain Information* Aalyss o Mult-product Break-eve wth Ucerta Iormato* Lazzar Lusa L. - Morñgo María Slva Facultad de Cecas Ecoómcas Uversdad de Bueos Ares 222 Córdoba Ave. 2 d loor C20AAQ Bueos Ares - Argeta lazzar@eco.uba.ar

More information

Aggregation Functions and Personal Utility Functions in General Insurance

Aggregation Functions and Personal Utility Functions in General Insurance Acta Polytechca Huarca Vol. 7, No. 4, 00 Areato Fuctos ad Persoal Utlty Fuctos Geeral Isurace Jaa Šprková Departmet of Quattatve Methods ad Iformato Systems, Faculty of Ecoomcs, Matej Bel Uversty Tajovského

More information

Regression Analysis. 1. Introduction

Regression Analysis. 1. Introduction . Itroducto Regresso aalyss s a statstcal methodology that utlzes the relato betwee two or more quattatve varables so that oe varable ca be predcted from the other, or others. Ths methodology s wdely used

More information

A Study of Unrelated Parallel-Machine Scheduling with Deteriorating Maintenance Activities to Minimize the Total Completion Time

A Study of Unrelated Parallel-Machine Scheduling with Deteriorating Maintenance Activities to Minimize the Total Completion Time Joural of Na Ka, Vol. 0, No., pp.5-9 (20) 5 A Study of Urelated Parallel-Mache Schedulg wth Deteroratg Mateace Actvtes to Mze the Total Copleto Te Suh-Jeq Yag, Ja-Yuar Guo, Hs-Tao Lee Departet of Idustral

More information

An Approach to Evaluating the Computer Network Security with Hesitant Fuzzy Information

An Approach to Evaluating the Computer Network Security with Hesitant Fuzzy Information A Approach to Evaluatg the Computer Network Securty wth Hestat Fuzzy Iformato Jafeg Dog A Approach to Evaluatg the Computer Network Securty wth Hestat Fuzzy Iformato Jafeg Dog, Frst ad Correspodg Author

More information

Chapter Eight. f : R R

Chapter Eight. f : R R Chapter Eght f : R R 8. Itroducto We shall ow tur our atteto to the very mportat specal case of fuctos that are real, or scalar, valued. These are sometmes called scalar felds. I the very, but mportat,

More information

MODELLING OF STOCK PRICES BY THE MARKOV CHAIN MONTE CARLO METHOD

MODELLING OF STOCK PRICES BY THE MARKOV CHAIN MONTE CARLO METHOD ISSN 8-80 (prt) ISSN 8-8038 (ole) INTELEKTINĖ EKONOMIKA INTELLECTUAL ECONOMICS 0, Vol. 5, No. (0), p. 44 56 MODELLING OF STOCK PRICES BY THE MARKOV CHAIN MONTE CARLO METHOD Matas LANDAUSKAS Kauas Uversty

More information

Optimization Model in Human Resource Management for Job Allocation in ICT Project

Optimization Model in Human Resource Management for Job Allocation in ICT Project Optmzato Model Huma Resource Maagemet for Job Allocato ICT Project Optmzato Model Huma Resource Maagemet for Job Allocato ICT Project Saghamtra Mohaty Malaya Kumar Nayak 2 2 Professor ad Head Research

More information

Speeding up k-means Clustering by Bootstrap Averaging

Speeding up k-means Clustering by Bootstrap Averaging Speedg up -meas Clusterg by Bootstrap Averagg Ia Davdso ad Ashw Satyaarayaa Computer Scece Dept, SUNY Albay, NY, USA,. {davdso, ashw}@cs.albay.edu Abstract K-meas clusterg s oe of the most popular clusterg

More information

A particle swarm optimization to vehicle routing problem with fuzzy demands

A particle swarm optimization to vehicle routing problem with fuzzy demands A partcle swarm optmzato to vehcle routg problem wth fuzzy demads Yag Peg, Ye-me Qa A partcle swarm optmzato to vehcle routg problem wth fuzzy demads Yag Peg 1,Ye-me Qa 1 School of computer ad formato

More information

We investigate a simple adaptive approach to optimizing seat protection levels in airline

We investigate a simple adaptive approach to optimizing seat protection levels in airline Reveue Maagemet Wthout Forecastg or Optmzato: A Adaptve Algorthm for Determg Arle Seat Protecto Levels Garrett va Ryz Jeff McGll Graduate School of Busess, Columba Uversty, New York, New York 10027 School

More information

10.5 Future Value and Present Value of a General Annuity Due

10.5 Future Value and Present Value of a General Annuity Due Chapter 10 Autes 371 5. Thomas leases a car worth $4,000 at.99% compouded mothly. He agrees to make 36 lease paymets of $330 each at the begg of every moth. What s the buyout prce (resdual value of the

More information

Report 05 Global Fixed Income

Report 05 Global Fixed Income Report 05 Global Fxed Icome From Dec 1999 to Dec 2014 31/12/1999 31 December 1999 31/12/2014 Rep05, Computed & Prted: 17/06/2015 11:24 New Performace Idcator (01/01/12) 100% Barclays Aggregate Global Credt

More information

Australian Climate Change Adaptation Network for Settlements and Infrastructure. Discussion Paper February 2010

Australian Climate Change Adaptation Network for Settlements and Infrastructure. Discussion Paper February 2010 Australa Clmate Chage Adaptato Network for Settlemets ad Ifrastructure Dscusso Paper February 2010 The corporato of ucertaty assocated wth clmate chage to frastructure vestmet apprasal Davd G. Carmchael

More information

Using Phase Swapping to Solve Load Phase Balancing by ADSCHNN in LV Distribution Network

Using Phase Swapping to Solve Load Phase Balancing by ADSCHNN in LV Distribution Network Iteratoal Joural of Cotrol ad Automato Vol.7, No.7 (204), pp.-4 http://dx.do.org/0.4257/jca.204.7.7.0 Usg Phase Swappg to Solve Load Phase Balacg by ADSCHNN LV Dstrbuto Network Chu-guo Fe ad Ru Wag College

More information

Banking (Early Repayment of Housing Loans) Order, 5762 2002 1

Banking (Early Repayment of Housing Loans) Order, 5762 2002 1 akg (Early Repaymet of Housg Loas) Order, 5762 2002 y vrtue of the power vested me uder Secto 3 of the akg Ordace 94 (hereafter, the Ordace ), followg cosultato wth the Commttee, ad wth the approval of

More information

Load Balancing Control for Parallel Systems

Load Balancing Control for Parallel Systems Proc IEEE Med Symposum o New drectos Cotrol ad Automato, Chaa (Grèce),994, pp66-73 Load Balacg Cotrol for Parallel Systems Jea-Claude Heet LAAS-CNRS, 7 aveue du Coloel Roche, 3077 Toulouse, Frace E-mal

More information

M. Salahi, F. Mehrdoust, F. Piri. CVaR Robust Mean-CVaR Portfolio Optimization

M. Salahi, F. Mehrdoust, F. Piri. CVaR Robust Mean-CVaR Portfolio Optimization M. Salah, F. Mehrdoust, F. Pr Uversty of Gula, Rasht, Ira CVaR Robust Mea-CVaR Portfolo Optmzato Abstract: Oe of the most mportat problems faced by every vestor s asset allocato. A vestor durg makg vestmet

More information

A DISTRIBUTED REPUTATION BROKER FRAMEWORK FOR WEB SERVICE APPLICATIONS

A DISTRIBUTED REPUTATION BROKER FRAMEWORK FOR WEB SERVICE APPLICATIONS L et al.: A Dstrbuted Reputato Broker Framework for Web Servce Applcatos A DISTRIBUTED REPUTATION BROKER FRAMEWORK FOR WEB SERVICE APPLICATIONS Kwe-Jay L Departmet of Electrcal Egeerg ad Computer Scece

More information

The paper presents Constant Rebalanced Portfolio first introduced by Thomas

The paper presents Constant Rebalanced Portfolio first introduced by Thomas Itroducto The paper presets Costat Rebalaced Portfolo frst troduced by Thomas Cover. There are several weakesses of ths approach. Oe s that t s extremely hard to fd the optmal weghts ad the secod weakess

More information

Automated Event Registration System in Corporation

Automated Event Registration System in Corporation teratoal Joural of Advaces Computer Scece ad Techology JACST), Vol., No., Pages : 0-0 0) Specal ssue of CACST 0 - Held durg 09-0 May, 0 Malaysa Automated Evet Regstrato System Corporato Zafer Al-Makhadmee

More information

Numerical Comparisons of Quality Control Charts for Variables

Numerical Comparisons of Quality Control Charts for Variables Global Vrtual Coferece Aprl, 8. - 2. 203 Nuercal Coparsos of Qualty Cotrol Charts for Varables J.F. Muñoz-Rosas, M.N. Pérez-Aróstegu Uversty of Graada Facultad de Cecas Ecoócas y Epresarales Graada, pa

More information

Security Analysis of RAPP: An RFID Authentication Protocol based on Permutation

Security Analysis of RAPP: An RFID Authentication Protocol based on Permutation Securty Aalyss of RAPP: A RFID Authetcato Protocol based o Permutato Wag Shao-hu,,, Ha Zhje,, Lu Sujua,, Che Da-we, {College of Computer, Najg Uversty of Posts ad Telecommucatos, Najg 004, Cha Jagsu Hgh

More information

Statistical Intrusion Detector with Instance-Based Learning

Statistical Intrusion Detector with Instance-Based Learning Iformatca 5 (00) xxx yyy Statstcal Itruso Detector wth Istace-Based Learg Iva Verdo, Boja Nova Faulteta za eletroteho raualštvo Uverza v Marboru Smetaova 7, 000 Marbor, Sloveja va.verdo@sol.et eywords:

More information

Beta. A Statistical Analysis of a Stock s Volatility. Courtney Wahlstrom. Iowa State University, Master of School Mathematics. Creative Component

Beta. A Statistical Analysis of a Stock s Volatility. Courtney Wahlstrom. Iowa State University, Master of School Mathematics. Creative Component Beta A Statstcal Aalyss of a Stock s Volatlty Courtey Wahlstrom Iowa State Uversty, Master of School Mathematcs Creatve Compoet Fall 008 Amy Froelch, Major Professor Heather Bolles, Commttee Member Travs

More information

Optimal Packetization Interval for VoIP Applications Over IEEE 802.16 Networks

Optimal Packetization Interval for VoIP Applications Over IEEE 802.16 Networks Optmal Packetzato Iterval for VoIP Applcatos Over IEEE 802.16 Networks Sheha Perera Harsha Srsea Krzysztof Pawlkowsk Departmet of Electrcal & Computer Egeerg Uversty of Caterbury New Zealad sheha@elec.caterbury.ac.z

More information

Curve Fitting and Solution of Equation

Curve Fitting and Solution of Equation UNIT V Curve Fttg ad Soluto of Equato 5. CURVE FITTING I ma braches of appled mathematcs ad egeerg sceces we come across epermets ad problems, whch volve two varables. For eample, t s kow that the speed

More information

Report 06 Global High Yield Bonds

Report 06 Global High Yield Bonds Rep06, Computed & Prted: 17/06/2015 11:25 Report 06 Global Hgh Yeld Bods From Dec 2000 to Dec 2014 31/12/2000 31 December 1999 31/12/2014 New Bechmark (01/01/13) 80% Barclays Euro HY Ex Facals 3% Capped

More information

Classic Problems at a Glance using the TVM Solver

Classic Problems at a Glance using the TVM Solver C H A P T E R 2 Classc Problems at a Glace usg the TVM Solver The table below llustrates the most commo types of classc face problems. The formulas are gve for each calculato. A bref troducto to usg the

More information

A Bayesian Networks in Intrusion Detection Systems

A Bayesian Networks in Intrusion Detection Systems Joural of Computer Scece 3 (5: 59-65, 007 ISSN 549-3636 007 Scece Publcatos A Bayesa Networs Itruso Detecto Systems M. Mehd, S. Zar, A. Aou ad M. Besebt Electrocs Departmet, Uversty of Blda, Algera Abstract:

More information

Common p-belief: The General Case

Common p-belief: The General Case GAMES AND ECONOMIC BEHAVIOR 8, 738 997 ARTICLE NO. GA97053 Commo p-belef: The Geeral Case Atsush Kaj* ad Stephe Morrs Departmet of Ecoomcs, Uersty of Pesylaa Receved February, 995 We develop belef operators

More information

On formula to compute primes and the n th prime

On formula to compute primes and the n th prime Joural's Ttle, Vol., 00, o., - O formula to compute prmes ad the th prme Issam Kaddoura Lebaese Iteratoal Uversty Faculty of Arts ad ceces, Lebao Emal: ssam.addoura@lu.edu.lb amh Abdul-Nab Lebaese Iteratoal

More information

An Evaluation of Naïve Bayesian Anti-Spam Filtering Techniques

An Evaluation of Naïve Bayesian Anti-Spam Filtering Techniques Proceedgs of the 2007 IEEE Workshop o Iformato Assurace Uted tates Mltary Academy, West Pot, Y 20-22 Jue 2007 A Evaluato of aïve Bayesa At-pam Flterg Techques Vkas P. Deshpade, Robert F. Erbacher, ad Chrs

More information