Comparison on Modelling the Relative. Risk Estimation: Bayesian Study
|
|
- Clare Shields
- 6 years ago
- Views:
Transcription
1 Appled Mathematcal Scences, Vol. 4, 00, no. 54, Comparson on Modellng the Relatve Rsk Estmaton: Bayesan Study Rafda M. Elobad and Noor Akma I. Laboratory of Appled & Computatonal Statstcs Insttute for Mathematcal Research Unversty Putra Malaysa, Serdang, Selangor, Malaysa Abstract The estmaton of the dsease ncdents was prevously analyzed usng a classcal approach. However, ths approach features large outlyng relatve rsks and consdered as msleadng due to several major problems. Some approaches such as the herarchcal Bayesan method have been adopted n the lterature n order to overcome these problems. The purpose of ths study s to compare between herarchcal Bayesan models that mprove the relatve rsk estmaton. The focus les on examnng the performance of dfferent sets of denstes va montorng the hstory graphs, estmatng the potental scale reducton factors and conductng senstvty analyss for dfferent choce of pror nformaton. The best model ft s accomplshed by conductng a goodness of ft test. The study s appled on Scotland lp cancer data set. The results show that for models wth large number of parameters, more teraton s needed to acheve the convergence. The study also shows that dagnostc test and senstvty analyss are mportant to decde about the stablty and the the nfluence of the choce of the pror denstes. The DIC results were n lne wth the prevous results and provde a good method of comparson. Mathematcs Subject Classfcaton: 6-07, 6P0
2 664 R. M. Elobad and N. Akma I. Keywords: herarchcal Bayesan models, relatve rsk, potental scale reducton, senstvty analyss, convergence, pror denstes, DIC. Introducton The Bayesan approach to statstcal analyss provdes a cohesve framework for combnng complex data models and external knowledge. The advantages n usng the Bayesan method justfy the ncreased efforts n computatons and pror determnaton [0]. The herarchcal Bayesan approach has become a standard n the epdemology and publc health lterature, see [, 8, 7,, 8]. Ths approach s called herarchcal because t uses multple level of analyss n an teratve way. Unlke the conventonal statstcal nference, whch derves the average estmates of the parameters, the herarchcal Bayesan modellng produces parameter estmates for each ndvdual analyss unt by borrowng nformaton from all analyss unts,.e. the customary Bayesan borrow of strength effect. As many herarchcal Bayesan models used n the lterature to nvestgate the relatve rsk estmaton, we lmt ths study to nvestgate three types of herarchcal Bayesan models. These models were classfed as follow. The nterregonal varablty model (IVM) whch ncludes the covarate effect and the nterregonal (local or non spatal) varablty effect, the correlated varablty model (CVM) that has the capablty of ncorporatng the covarate effect and the effect of the rsk n the adjacent countes (the spatal varablty) and the global spatal model (GSM) that accommodates the covarate, nterregonal and the spatal effects. The three models were ntroduced n such away to produce stable, accurate and smooth relatve rsk estmaton. In ths study, the relatve rsk n each county assumed to depend on the adjacent neghbours [9, 3, 5]. The MCMC smulaton was used to estmate the dfferent sources that caused the relatve rsk varablty n each county usng the three Bayesan selected models. However, the selected models have dfferent nfluence n the relatve rsk estmaton. Ths usually depends on the choce of the pror denstes, the number of teraton used n the smulaton, the number of parameter n each model...etc. Ths study ntroduces dfferent methods to compare between the selected herarchcal Bayesan models.. Method. Estmatng relatve rsks usng the IVM, CVM and GSM Several factors affectng the relatve rsks among the countes wthn the lattce grd lead to relatve rsk heterogenety. In ths study, two man sources were dentfed as
3 Comparson on modellng the relatve rsk estmaton 665 the causes of ths heterogenety, the covarate effect whch nfluences the occurrences of the dsease n the dfferent countes and the effect of the over-dsperson [5,, 0] between the countes. The over-dsperson was separated nto two man effects, the nterregonal varablty due to dfferent envronmental (local, natural, or bult) factors, and the spatal varablty whch s due to the spatal autocorrelaton among the dfferent countes [4]. The dsease relatve rsk was estmated usng the standardzed mortalty rates (SMR). In the presence of the prevous factors, three herarchcal models IVM, CVM and GSM were used. Two stages were consdered to buld all the three models. In the frst stage, and snce rare events were dealt wth n buldng the herarchcal Bayesan models, the Posson lkelhood was specfed for the observed dsease ncdents, gven the vector of the dsease relatve rsk. In the second stage, a set of pror denstes for each model was selected over the space of the possble relatve rsk parameters and hyper-parameters to get a set of posteror means for the relatve rsks, gven the observed dsease ncdences. The constructon method for each model s elaborated as follow. For the IVM, CVM and GSM the SMR s defned as n the followng equatons respectvely, ˆ X ˆ SMR e β + ψ β + = = θ, () ˆ X ˆ SMR e β + ψ β + = =, () ψˆ = SMR ˆ X e β + β + θ + =, (3) where ψ s the relatve rsk parameter n regon whch estmated usng the standardzed mortalty rate ( SMR ˆ ), β s the ntercept term, β s the coeffcent for the covarate X, θ s the random effect representng the nterregonal varablty, and s the spatal varablty random effect n countes =,,.,N. Pror denstes were specfed for each parameter n each model. As ponted out by Kelsall and Wakefeld [9], care s needed when choosng the pror denstes for the parameters of nterest because the posteror denstes of the parameters can be senstve. In the IVM, the nterregonal varablty θ was assumed as an ndependent gamma pror, condtonal on the hyper-parameters α and α. Wthout the constant term, the gamma pror densty was proportonal to the followng: f(, ) α θ α α θ e α θ, (4) where α α s the pror expectaton for θ, and α α s the pror varance. Wthout pror expectatons about the drecton and magntude of the covarate effects
4 666 R. M. Elobad and N. Akma I. and condtonal on known hyper-parameters, the ntercept β and the covarate coeffcent β are assumed to have vague but nformatve normal pror, ( a) b f( β a, b) e β, (5) f( β r, d) e β ( r) d. (6) where abr,, and d are fxed known values. The hyper-parameters α and α of the gamma densty are also gven prors p( α Y ) and p( α Y ). The prors selected was the exponental dstrbuton wth constant second order hyper-parameters c and m for each α and α, respectvely, that s: c f ( α c) e α, (7) m f( α m) e α. (8) For the CVM, condtonal on the hyper-parameter τ whch controls the varablty of, the spatal varablty was modelled as ntrnsc condtonal autoregressve (ICAR) model [7]. The most common densty of ths pror [] has the jont dstrbuton proportonal to: τ n / τ / τ exp ( ) τ exp ω ( ), (9) = where τ = σ s the precson term, = ω denotes the average of the neghbourng that s adjacent to, and ω s the number of the adjacent countes. The vague flat pror dstrbuton was used for the ntercept term β to gve the desred propertes of the mproper ICAR pror [6]. The covarate coeffcent β was assumed to have a normal pror dstrbuton, as n Equaton (6) wth fxed hyper-parameters r and d. The vague Gamma dstrbuton was chosen as a pror densty p( τ Y ) for the hyper-parameter τ of the spatal varablty, wth mean l l and varance l l. Wth no pror estmaton for the precsons of the random effects, small values of l, l were chosen to assume the large varance. The pror for precson hyper-parameter τ s proportonal to the followng densty, l f ( τ l, l ) τ e l τ. (0)
5 Comparson on modellng the relatve rsk estmaton 667 For the thrd model (the GSM), condtonal on the precson hyper-parameter τ θ, the nterregonal varablty θ s assumed as a conjugate normal ndependent pror, that s: τθ τθ f ( θ τθ ) = exp( θ ), () π where τθ = σθ s the precson term, whch s the nverse of the nterregonal varance. Condtonal on the hyper-parameter τ, the spatal varablty s modelled as the ntrnsc condtonal autoregressve (ICAR) model, where has a normal dstrbuton as gven n Equaton (9). The pror denstes for the fx parameters β and β were represented as n the CVM. Prors p( τ θ Y ) and p( τ Y ) were chosen for the hyper-parameters, τ θ and τ of the nterregonal and spatal varablty. The vague Gamma dstrbuton wth mean k k and varance k k was chosen for τ θ as follow, k kτθ f( τθ k, k) τθ e. () The gamma pror defned n Equaton (0) was chosen for the spatal precson term τ. Small values for k, k, l and l were chosen to assume large varances, as there were no pror nformaton for precsons of the random effects.. The smulaton study The IVM smulaton: no pror expectaton about the ntercept β and the covarate effects β were avalable, hence a large value was chosen for the fxed hyperparameters whch represented the varances b and d of the pror normal denstes of β and β, respectvely ( b= d = 00 ). For the pror exponental densty of the hyperparameters α and α, small values for the second order hyper-parameters, c= m= 0.0, were chosen to assume large varances, c and α, m for α and respectvely. Two thousand teratons, wth the frst 500 teratons dscarded from each chan as pre-convergence burn n, were used to start the smulaton. The CVM smulaton: to carry out the posteror analyss, fxed values were selected n such a way that produces vague denstes whch can be used to obtan the posteror denstes for the parameters. A large value was chosen for the varance (d=000) of the pror normal dstrbuton of the covarate term β. Wth no pror estmaton for the precson of the random effect termτ, the fx values of 0.5 and were chosen for the second order hyper-parameters of Gamma pror densty l, and l, respectvely.
6 668 R. M. Elobad and N. Akma I. Usng the MCMC method, more teratons were expected to be needed n order to start the analyss, due to the spatal structure whch has been ncluded n the model. Hence, the smulaton started wth three thousand smulatons from the posteror dstrbuton wth the frst 500 teratons dscarded from the results. The GSM smulaton: large value for the varance (d = 0000) was selected for the pror densty of β to produce vague densty, whle vague but proper pror dstrbutons Gamma (0.5, 0.05) and Gamma (0.5, 0.0) were adopted for the precson of the random effects τ θ andτ, respectvely. The analyss was started wth three thousand smulatons from the posteror dstrbuton..3 Models evaluaton To nvestgate the performance of the three models, the analyss was appled to the lp cancer data n 56 countes n Scotland over the perod 973 to 980. The data ncluded the name and the number of 56 dstrcts n Scotland, the observed number of male lp cancer cases n each county Y, the number of male populaton n, a covarate X measurng the percentage of the county s populaton engaged n agrculture, forestry, and fshng (AFF), and the poston of each county expressed as a lst of adjacent countes. Ths data had been orgnally analyzed by Clayton and Kaldor [4] and Breslow and Clayton [6]. Snce the lp cancer dsease s a rare and non-contagous, the number of the observed ncdents s assumed to be mutually ndependent, and to follow Posson dstrbuton. The sets of pror denstes, n addton to the Posson lkelhood are used to obtan the jont posteror densty and consequently the condtonal posteror densty for each parameter n each model. The convergence of the samplng process used to estmate the parameters was assessed by montorng the trace plots of the parameters. To acheve the convergence two parallel chans were used for each model to ensure a complete coverage of the sample space. Sutable number of teraton was also used after takng nto consderaton the number and the expresson of the parameters, the sample sze and the conjugacy of the pror nformaton for each model. Convergence dagnostc was also appled by estmatng the potental scale reducton factor R ) for each parameter of nterest []. The unty value of R ) ndcates that the convergence s reached; otherwse the smulaton can be contnuously run by ncreasng the number of teraton per chan. The precson of the pror nformaton was assessed usng senstvty analyss for each model. The senstvty analyss was conducted n order to frst, nvestgate the nfluence of the choce of the pror denstes n estmatng the relatve rsk, and second, to nvestgate whether the results n the analyss remaned essentally unchanged n the presence of dfferent pror nformaton. Once the models performance was assessed, the study made use of the development of the MCMC methods, whch had made t possble to ft and compare
7 Comparson on modellng the relatve rsk estmaton 669 ncreasngly large classes of models [3]. We used the devance nformaton crteron (DIC), whch was ntroduced by Spegelhater et al. [3] to compare between the herarchcal Bayesan models. The DIC s represented by the followng equaton; DIC = D + p = D D( Θ ), (3) D where D s the posteror expectaton of the devance, pd s the effectve number of parameters and D( Θ ) s the devance evaluated at the posteror expectaton. As wth all penalzed lkelhood crtera, the DIC conssts of two terms; one represents the goodness of ft, whle the other, a penalty for ncreasng the complexty of the model. One of the man reasons, whch encouraged the use of ths method n makng a comparson between the proposed Bayesan models, s that, besdes ts generalty, t may readly be calculated n the MCMC process. 3. Results and Dscusson The convergence of the MCMC smulaton to the posteror dstrbutons for the selected parameters was observed usng the WnBUGS software. Two parallel samplng chans wth two dfferent sets of ntal values for each model were used. Fgures (), () and (3) llustrate the hstory graphs for selected posteror means of the SMR and selected parameters for the IVM, CVM and GSM, respectvely.
8 670 R. M. Elobad and N. Akma I. (a) (b) Fgure. Hstory graphs for selected a) posteror means of the SMR and b) posteror parameters usng the IVM.
9 Comparson on modellng the relatve rsk estmaton 67 (a) (b) Fgure. Hstory graphs for selected a) posteror means of the SMR and b) posteror parameters usng the CVM. The hstory graphs n Fgure () show the output from the two parallel Gbbs samplng chans usng the IVM. Fgure (a) ndcates that the posteror means of the estmated SMR have converged well at 000 and 000 teratons, respectvely; whle Fgure (b) shows that the posteror means for some parameters such as α are not acheved at 000 teratons. By montorng the convergence n the trace plots of the CVM, the prmary results observed from the hstory graphs n Fgures (a) and (b) showed that the convergence had been reached for all the parameters. However, the trace plots of the parameters n the GSM show that the convergence has not been reached for some
10 67 R. M. Elobad and N. Akma I. parameters wth less than sx thousand teratons. Ths s due to the large number of parameters n the model. Thus, three thousand more teratons have been updated. The hstory graphs for the selected posteror means of the SMR and selected parameters are dsplayed n Fgures (3a) and (3b), respectvely. (a) (b) Fgure 3. Hstory graphs for selected a) posteror means of the SMR and b) posteror parameters usng the GSM. To confrm that the convergence of all the parameters had been acheved, the potental scale reducton factor R ) was estmated for each parameter of nterest. In the IVM, Ths factor showed that the convergence had not been acheved for some parameters by usng only 000 teratons,.e. the value of R ) s not equal to one. Ths was depcted n Fgures (4a) and (4b) whch show the overlap of the two chans, R )
11 Comparson on modellng the relatve rsk estmaton 673 values and the posteror nferences for each estmator usng the IVM. Fgures (4a) and (4b) depct the output from 000 and 000 MCMC smulatons, respectvely. The results n these two fgures show obvous results of non-convergence for some parameters than that obtaned by the hstory graphs (Fgure ). Another 000 teratons were carred out n order to acheve the convergence. Usng a total of 3000 teratons ndcated that the potental scale reducton factor was between and. for all the parameters as dsplayed n Fgure (4c). From the fgure, the R ) values confrm that the convergence has been acheved for all the parameters. (a) (b)
12 674 R. M. Elobad and N. Akma I. (c) Fgure 4. Inference for the IVM ncludng the R ) Estmaton from a) 000 b) 000 and c) 3000 MCMC smulatons. Fgure 5. Inference for the CVM ncludng the R ) estmaton from 3000 MCMC smulatons.
13 Comparson on modellng the relatve rsk estmaton 675 Fgure 6. Inference for the GSM ncludng the R ) estmaton from 6000 MCMC smulaton. The potental scale reducton factor R ) was re-estmated for the parameters usng 3000 MCMC smulaton for CVM and 6000 MCMC smulatons for GSM. It was found that the convergence had been acheved for all the parameters. Fgures (5) and (6) show that R ) values are equal to the unty value, suggestng that the models converge well after several thousand teratons. Ths result ndcated a good agreement wth the hstory graphs results (Fgures and 3 for CVM and GSM, respectvely) and show that no more teratons are needed. The senstvty analyss was carred out by conductng several trals to dfferent choces of fxed values for the hyper-parameters abr,, and d, and the second order hyper-parameters c and m n the IVM. The MCMC smulaton was carred out for each tral, and the samples of the posterors were summarzed. Later, these samples were checked to fnd out whether they had mposed a practcal mpact on the nterpretatons or decsons made. All the trals gave almost dentcal results to the results shown n Fgure (7). The fgure dsplays the relatve rsk estmaton for the 56 countes of Scotland. Overall, the mean of the relatve rsk s.45, whle the standard devaton s 0.9. Another senstvty analyss was conducted usng dfferent pror nformaton n CVM. The posteror quanttes of nterest were recomputed by specfyng vague but proper pror dstrbutons, Normal (0, 00) and Gamma (0.5, 0.05) for the covarate and the precson parameters, respectvely; followed by other trals wth dfferent sets
14 [7] [8] [9] [0][] [][3] 676 R. M. Elobad and N. Akma I. of fxed values for both prors. The output n all the tral remaned almost the same for all the parameters, ndcatng the robustness of the results wth dfferent choces of pror nformaton. The overall mean of the relatve rsk usng the CVM s.43, as shown n the box plot of the estmated SMR n Fgure (8). The standard devaton has also been reduced to.06, as compared to the results derved from the IVM. countes vs SMRhat 0.0 [] [] [3] [4] [5] 5.0 [6] [4] [5][6] [7] [8] [9] [0] [] [3] [][3] [4] [5][6] [7] [8][9][30] [3] [33] [39] [34][35][36][37][38] [4] [43] [55] [56] [40] [46][47] [5][5] [4] [44] [48][49] [53][54] [45] [50] 0.0 Countes Fgure 7. Box plots of the IVM posteror estmaton of the SMR for Scotland countes. countes vs SMRhat 0.0 [] [] [3] [] 5.0 [4] [5] [7] [0] [3] [6] [9] [4] [5] [6][7] [9] [0] [3] [8] [] [8] [][] [3] [5] [9] [6][7][8] [39] [43] [33] [4] [30][3] [34] [35][36][37] [55] [56] [4] [38] [40] [46] [4] [44] [47][48][49] [50][5][5] [45] [53][54] 0.0 countes Fgure 8. Box plots of the CVM posteror estmaton of SMR for Scotland countes.
15 Comparson on modellng the relatve rsk estmaton 677 countes vs SMRhat 0.0 [] [] [3] [] 5.0 [4] [5] [6] [7] [0] [9] [] [8] [3] [4] [7] [6] [5] [9] [3] [8] [0] [][][3] [4] [5][6][7][8][9] [30][3] [33] [39] [43] [34][35][36] [37] [55] [56] [38] [40] [4] [4] [46] [44] [47][48] [45] [49] [50][5][5] [53][54] 0.0 countes Fgure 9. Box plots of the GSM posteror estmaton of the SMR for Scotland countes. In the GSM the senstvty analyss ndcates that the pror dstrbutons gave almost dentcal results, suggestng that the results are robust to changes n pror nformaton. The GSM was found to produce shrnkage n the estmaton of the posteror relatve rsk towards the overall mean (.456), as shown n Fgure (9). The standard devaton of the dfferent trals of the relatve rsk estmaton showed the smallest value (0.89) among the Bayesan models. Ths confrms that the pror used for ths model are more robust and accurate comparng wth the other two models. The DIC was obtaned to compare between the models. The selecton was done based on the DIC values computed for each model, usng the MCMC method. Model wth a smaller DIC ndcates a better fttng model. Table () lsts the devance summares for each model, the posteror mean of the devance, D, the devance evaluated at the posteror mean of the parameters, D( Θ ), the effectve number of the parameters p D and the DIC. The smaller values of D and D( Θ ) showed that models wth spatal random effect (CVM and GSM) have a better average ft to data than the model wth the nterregonal varablty (IVM). The results also show that the best fttng s acheved when the full model (GSM) s used due to ts small D and D( Θ ) values. Ths partcular model also has a smaller DIC value (306.69) and extra effectve parameter, p D = As a result, t s suggested that ths s the most robust and approprate model to be used.
16 678 R. M. Elobad and N. Akma I. Table. Devance summares for the herarchcal Bayesan models. Model D D( Θ ) p D DIC IVM CVM GSM Concluson The work n ths paper had prmarly been concerned wth producng stable estmaton for the relatve rsk assocated wth the spread of the dsease n spatally arranged regons. The purpose s to provde dfferent methods to assess the performance of the herarchcal Bayesan models IVM, CVM and GSM. These models nvestgate the covarate and over-dsperson effects that mght affect the relatve rsk estmaton. The study was carred out to choose the most approprate model whch could ft the data well. The Posson lkelhood and dfferent sets of the pror denstes were used to obtan the jont posteror densty, followed by the condtonal posteror densty for each parameter. The study dscussed the results of fttng these models to the Scotland lp cancer data. An MCMC smulaton was conducted usng two chans wth selected ntal values and specfc number of teratons for each model. Burn-n perods were also used to confrm a statonary convergence whch s ndependent of the ntal parameter values for each model. The convergences were assessed for each model by montorng the trace plots whch summarzed n hstory graphs for all the parameters of nterest. Meanwhle, the results from the potental reducton factor R ) show more obvous results of nonconvergence for some parameters than that obtaned by the hstory graphs n IVM. Ths ndcates the mportant of applyng dagnostc test to examne the convergence. Increasng the number of teraton n the IVM and n the subsequent models showed that the R ) values vary between and. ndcatng that the convergence was acheved. A senstvty analyss was also conducted n order to study the nfluence of the choce of the pror denstes. Ths analyss nvestgated the adequacy of the dfferent pror nformaton used n each herarchcal model. The results show that the GSM provde more adequate pror denstes compared to the IVM and CVM. Ths was also confrmed when adoptng the DIC to compare between the nested models. The DIC results showed that the GSM had smaller value and extra effectve parameters. Ths
17 Comparson on modellng the relatve rsk estmaton 679 suggested that the GSM could perform well as compared to the other herarchcal models. Ths result also ndcates a good agreement wth the results obtaned by Banerjee et al. []. The present study show that wthn the Bayesan approach the dfferent methods used to demonstrate the models provde good compromse between satsfactory and approprateness for the data of nterest. The model wth suffcent characterstcs wll mprove the relatve rsk estmaton and help the researcher to plan and evaluate strateges used to prevent llness. REFERENCES [] A. Gelman, J.B. Carlne, H.S. Stern and D.B Rubn, Bayesan Data Analyss, st ed. Chapman and Hall Text, In Statstcal Scence Seres, Chapman & Hall, London, 995. [] B.P. Carln and T.A. Lous, Bayes and Emprcal Bayes Methods for Data Analyss, Chapman and Hall/CRC, New York, 000. [3] D. Spegelhalter, N. Best, B.P. Carln and A. van dn Lnde, Bayesan measures of model complexty and ft (wth dscusson), J. Roy. Statst. Soc. Ser. B, 64 (00), pp [4] D.G. Clayton and J. Kaldor, Emprcal Bayes estmates of age-standardzed relatve rsks for use n dsease mappng, Bometrcs, 43(987), pp [5] D.G. Clayton, A Monte Carlo method for Bayesan nference n fralty models, Bometrcs, 47 (99), pp [6] J. Besag and C. Kooperberg, On condtonal and ntrnsc auto regressons. Bomrtrka, 8(4) (995), pp [7] J. Besag, J. York and A. Molle, Bayesan mage restoraton, wth two applcatons n spatal statstc, Ann. Inst. Stat. Math., 43 (99), pp. -0. [8] J. Gll, Bayesan Methods for the Socal and Behavoural Scences, Chapman & Hall, New York, 00. [9] J.E. Kelsall and J.C. Wakefeld, Dscusson of Bayesan models for spatally correlated dsease and exposure data, by Best et al., In Bayesan Statstcs 6, ed., J.M. Bernardo, J.O. Berger, A.P. Davd, A.F.M Smth, Oxford Unversty Press, Oxford, 999, pp. 5.
18 680 R. M. Elobad and N. Akma I. [0] J.O. Berger, Statstcal Decson Theory and Bayesan Analyss, nd ed. Sprnger- Verlag, New York, 985. [] K. Manton, M. Woodbury and E. Stallard, A varance components approach to categorcal data models wth heterogeneous mortalty rates n North Carolna countes. Bometrcs, 37 (98), pp [] L. Bernardnell and C. Montomol, Emprcal Bayes versus fully Bayesan analyss of geographcal varaton of dsease rsk, Statst. Med. (99), pp [3] M.D. Ecker and A.E. Gelfand, Bayesan varograme modellng for an sotropc spatal process, J. Agrc. Bol. Envron. Statst, (997), pp [4] M.E. Rafda, Herarchcal Bayesan Spatal Models for Dsease Mortalty Rates, Ph.D. Dssertaton, Insttute for Mathematcal Research, UPM, 009. [5] M.J. Karson, M. Gaudard, E. Lnder and D. Snha, Bayesan analyss and computatons for spatal predctons (wth dscusson), Envronmental and Ecologcal Statstcs, 6 (999), pp [6] N.E. Breslow and D.G. Clayton, Approxmate nference n generalzed lnear mxed models, J. Amer. Statst. Assoc., 88 (993), pp [7] P. Congdon, Appled Bayesan Modellng, Wley, Chchester, 003. [8] P. Congdon, Bayesan Statstcal Modellng, Wley, Chchester, 00. [9] P.J. Dggle, J.A. Tawn and R.A. Moyeed, Model-based geostatstcs (wth dscusson), J. Roy. Statst. Soc., Ser. C (Appled Statstcs), 47 (998), pp [0] R. Tsutakawa, Mxed model for analyzng geographc varablty n mortalty rates, J. Amer. Stat. Assoc., 83 (988), pp [] S. Banerjee, B.P. Carln, and A.E. Gelfand, Herarchcal Modelng and Analyss for Spatal Data, Chapman and Hall/CRC Press, 004. [] T. Banerjee and B. Atanu, A new formulaton of stress strength relablty n a regresson setup, Journal of Statstcal Plannng and Inference, (-)(003), pp
19 Comparson on modellng the relatve rsk estmaton 68 [3] W.R. Glks, S. Rchardson and D.J. Spegelhalter, Markov Chan Monte Carlo n Practce, Chapman and Hall, New York, 996. Receved: November, 009
How its computed. y outcome data λ parameters hyperparameters. where P denotes the Laplace approximation. k i k k. Andrew B Lawson 2013
Andrew Lawson MUSC INLA INLA s a relatvely new tool that can be used to approxmate posteror dstrbutons n Bayesan models INLA stands for ntegrated Nested Laplace Approxmaton The approxmaton has been known
More informationSpatial Frailty Survival Model for Infection of Tuberculosis among HIV Infected Individuals
SSRG Internatonal Journal of Medcal Scence (SSRG-IJMS) Volume 4 Issue 2 February 217 Spatal Fralty Survval Model for Infecton of uberculoss among HIV Infected Indvduals Srnvasan R 1, Ponnuraja C 1, Rajendran
More informationEcon107 Applied Econometrics Topic 3: Classical Model (Studenmund, Chapter 4)
I. Classcal Assumptons Econ7 Appled Econometrcs Topc 3: Classcal Model (Studenmund, Chapter 4) We have defned OLS and studed some algebrac propertes of OLS. In ths topc we wll study statstcal propertes
More informationChapter 11: Simple Linear Regression and Correlation
Chapter 11: Smple Lnear Regresson and Correlaton 11-1 Emprcal Models 11-2 Smple Lnear Regresson 11-3 Propertes of the Least Squares Estmators 11-4 Hypothess Test n Smple Lnear Regresson 11-4.1 Use of t-tests
More informationComparison of Regression Lines
STATGRAPHICS Rev. 9/13/2013 Comparson of Regresson Lnes Summary... 1 Data Input... 3 Analyss Summary... 4 Plot of Ftted Model... 6 Condtonal Sums of Squares... 6 Analyss Optons... 7 Forecasts... 8 Confdence
More informationPsychology 282 Lecture #24 Outline Regression Diagnostics: Outliers
Psychology 282 Lecture #24 Outlne Regresson Dagnostcs: Outlers In an earler lecture we studed the statstcal assumptons underlyng the regresson model, ncludng the followng ponts: Formal statement of assumptons.
More informationLINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity
LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 31 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 6. Rdge regresson The OLSE s the best lnear unbased
More informationDurban Watson for Testing the Lack-of-Fit of Polynomial Regression Models without Replications
Durban Watson for Testng the Lack-of-Ft of Polynomal Regresson Models wthout Replcatons Ruba A. Alyaf, Maha A. Omar, Abdullah A. Al-Shha ralyaf@ksu.edu.sa, maomar@ksu.edu.sa, aalshha@ksu.edu.sa Department
More informationNegative Binomial Regression
STATGRAPHICS Rev. 9/16/2013 Negatve Bnomal Regresson Summary... 1 Data Input... 3 Statstcal Model... 3 Analyss Summary... 4 Analyss Optons... 7 Plot of Ftted Model... 8 Observed Versus Predcted... 10 Predctons...
More informationx = , so that calculated
Stat 4, secton Sngle Factor ANOVA notes by Tm Plachowsk n chapter 8 we conducted hypothess tests n whch we compared a sngle sample s mean or proporton to some hypotheszed value Chapter 9 expanded ths to
More informationNUMERICAL DIFFERENTIATION
NUMERICAL DIFFERENTIATION 1 Introducton Dfferentaton s a method to compute the rate at whch a dependent output y changes wth respect to the change n the ndependent nput x. Ths rate of change s called the
More informationStatistics for Economics & Business
Statstcs for Economcs & Busness Smple Lnear Regresson Learnng Objectves In ths chapter, you learn: How to use regresson analyss to predct the value of a dependent varable based on an ndependent varable
More informationBOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS. M. Krishna Reddy, B. Naveen Kumar and Y. Ramu
BOOTSTRAP METHOD FOR TESTING OF EQUALITY OF SEVERAL MEANS M. Krshna Reddy, B. Naveen Kumar and Y. Ramu Department of Statstcs, Osmana Unversty, Hyderabad -500 007, Inda. nanbyrozu@gmal.com, ramu0@gmal.com
More informationNon-Mixture Cure Model for Interval Censored Data: Simulation Study ABSTRACT
Malaysan Journal of Mathematcal Scences 8(S): 37-44 (2014) Specal Issue: Internatonal Conference on Mathematcal Scences and Statstcs 2013 (ICMSS2013) MALAYSIAN JOURNAL OF MATHEMATICAL SCIENCES Journal
More informationComparison of the Population Variance Estimators. of 2-Parameter Exponential Distribution Based on. Multiple Criteria Decision Making Method
Appled Mathematcal Scences, Vol. 7, 0, no. 47, 07-0 HIARI Ltd, www.m-hkar.com Comparson of the Populaton Varance Estmators of -Parameter Exponental Dstrbuton Based on Multple Crtera Decson Makng Method
More informationChapter 13: Multiple Regression
Chapter 13: Multple Regresson 13.1 Developng the multple-regresson Model The general model can be descrbed as: It smplfes for two ndependent varables: The sample ft parameter b 0, b 1, and b are used to
More informationLINEAR REGRESSION ANALYSIS. MODULE IX Lecture Multicollinearity
LINEAR REGRESSION ANALYSIS MODULE IX Lecture - 30 Multcollnearty Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur 2 Remedes for multcollnearty Varous technques have
More informationParametric fractional imputation for missing data analysis. Jae Kwang Kim Survey Working Group Seminar March 29, 2010
Parametrc fractonal mputaton for mssng data analyss Jae Kwang Km Survey Workng Group Semnar March 29, 2010 1 Outlne Introducton Proposed method Fractonal mputaton Approxmaton Varance estmaton Multple mputaton
More informationDr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur
Analyss of Varance and Desgn of Experment-I MODULE VII LECTURE - 3 ANALYSIS OF COVARIANCE Dr Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Any scentfc experment s performed
More informationDepartment of Quantitative Methods & Information Systems. Time Series and Their Components QMIS 320. Chapter 6
Department of Quanttatve Methods & Informaton Systems Tme Seres and Ther Components QMIS 30 Chapter 6 Fall 00 Dr. Mohammad Zanal These sldes were modfed from ther orgnal source for educatonal purpose only.
More information1. Inference on Regression Parameters a. Finding Mean, s.d and covariance amongst estimates. 2. Confidence Intervals and Working Hotelling Bands
Content. Inference on Regresson Parameters a. Fndng Mean, s.d and covarance amongst estmates.. Confdence Intervals and Workng Hotellng Bands 3. Cochran s Theorem 4. General Lnear Testng 5. Measures of
More informationStatistics for Managers Using Microsoft Excel/SPSS Chapter 13 The Simple Linear Regression Model and Correlation
Statstcs for Managers Usng Mcrosoft Excel/SPSS Chapter 13 The Smple Lnear Regresson Model and Correlaton 1999 Prentce-Hall, Inc. Chap. 13-1 Chapter Topcs Types of Regresson Models Determnng the Smple Lnear
More informationGlobal Sensitivity. Tuesday 20 th February, 2018
Global Senstvty Tuesday 2 th February, 28 ) Local Senstvty Most senstvty analyses [] are based on local estmates of senstvty, typcally by expandng the response n a Taylor seres about some specfc values
More informationUncertainty as the Overlap of Alternate Conditional Distributions
Uncertanty as the Overlap of Alternate Condtonal Dstrbutons Olena Babak and Clayton V. Deutsch Centre for Computatonal Geostatstcs Department of Cvl & Envronmental Engneerng Unversty of Alberta An mportant
More informationGENERALIZED LINEAR MIXED MODELS WITH SPATIAL RANDOM EFFECTS FOR SPATIO-TEMPORAL DATA: AN APPLICATION TO DENGUE FEVER MAPPING
Journal of Mathematcs and Statstcs, 9 (2): 137-143, 2013 ISSN 1549-3644 2013 do:10.3844/jmssp.2013.137.143 Publshed Onlne 9 (2) 2013 (http://www.thescpub.com/jmss.toc) GENERALIZED LINEAR MIXED MODELS WITH
More informationSpatial Modelling of Peak Frequencies of Brain Signals
Malaysan Journal of Mathematcal Scences 3(1): 13-6 (9) Spatal Modellng of Peak Frequences of Bran Sgnals 1 Mahendran Shtan, Hernando Ombao, 1 Kok We Lng 1 Department of Mathematcs, Faculty of Scence, and
More informationOn an Extension of Stochastic Approximation EM Algorithm for Incomplete Data Problems. Vahid Tadayon 1
On an Extenson of Stochastc Approxmaton EM Algorthm for Incomplete Data Problems Vahd Tadayon Abstract: The Stochastc Approxmaton EM (SAEM algorthm, a varant stochastc approxmaton of EM, s a versatle tool
More informationChapter 9: Statistical Inference and the Relationship between Two Variables
Chapter 9: Statstcal Inference and the Relatonshp between Two Varables Key Words The Regresson Model The Sample Regresson Equaton The Pearson Correlaton Coeffcent Learnng Outcomes After studyng ths chapter,
More informationISSN X Robust bayesian inference of generalized Pareto distribution
Afrka Statstka Vol. 112), 2016, pages 1061 1074. DOI: http://dx.do.org/10.16929/as/2016.1061.92 Afrka Statstka ISSN 2316-090X Robust bayesan nference of generalzed Pareto dstrbuton Fatha Mokran 1, Hocne
More informationBayesian predictive Configural Frequency Analysis
Psychologcal Test and Assessment Modelng, Volume 54, 2012 (3), 285-292 Bayesan predctve Confgural Frequency Analyss Eduardo Gutérrez-Peña 1 Abstract Confgural Frequency Analyss s a method for cell-wse
More informationCOMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS
Avalable onlne at http://sck.org J. Math. Comput. Sc. 3 (3), No., 6-3 ISSN: 97-537 COMPARISON OF SOME RELIABILITY CHARACTERISTICS BETWEEN REDUNDANT SYSTEMS REQUIRING SUPPORTING UNITS FOR THEIR OPERATIONS
More informationSemiparametric geographically weighted generalised linear modelling in GWR 4.0
Semparametrc geographcally weghted generalsed lnear modellng n GWR 4.0 T. Nakaya 1, A. S. Fotherngham 2, M. Charlton 2, C. Brunsdon 3 1 Department of Geography, Rtsumekan Unversty, 56-1 Tojn-kta-mach,
More informationConjugacy and the Exponential Family
CS281B/Stat241B: Advanced Topcs n Learnng & Decson Makng Conjugacy and the Exponental Famly Lecturer: Mchael I. Jordan Scrbes: Bran Mlch 1 Conjugacy In the prevous lecture, we saw conjugate prors for the
More informationA Robust Method for Calculating the Correlation Coefficient
A Robust Method for Calculatng the Correlaton Coeffcent E.B. Nven and C. V. Deutsch Relatonshps between prmary and secondary data are frequently quantfed usng the correlaton coeffcent; however, the tradtonal
More informationOn Outlier Robust Small Area Mean Estimate Based on Prediction of Empirical Distribution Function
On Outler Robust Small Area Mean Estmate Based on Predcton of Emprcal Dstrbuton Functon Payam Mokhtaran Natonal Insttute of Appled Statstcs Research Australa Unversty of Wollongong Small Area Estmaton
More informationMarkov Chain Monte Carlo Lecture 6
where (x 1,..., x N ) X N, N s called the populaton sze, f(x) f (x) for at least one {1, 2,..., N}, and those dfferent from f(x) are called the tral dstrbutons n terms of mportance samplng. Dfferent ways
More informationDr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur
Analyss of Varance and Desgn of Experment-I MODULE VIII LECTURE - 34 ANALYSIS OF VARIANCE IN RANDOM-EFFECTS MODEL AND MIXED-EFFECTS EFFECTS MODEL Dr Shalabh Department of Mathematcs and Statstcs Indan
More informationComputation of Higher Order Moments from Two Multinomial Overdispersion Likelihood Models
Computaton of Hgher Order Moments from Two Multnomal Overdsperson Lkelhood Models BY J. T. NEWCOMER, N. K. NEERCHAL Department of Mathematcs and Statstcs, Unversty of Maryland, Baltmore County, Baltmore,
More informationGeneralized Linear Methods
Generalzed Lnear Methods 1 Introducton In the Ensemble Methods the general dea s that usng a combnaton of several weak learner one could make a better learner. More formally, assume that we have a set
More informationKernel Methods and SVMs Extension
Kernel Methods and SVMs Extenson The purpose of ths document s to revew materal covered n Machne Learnng 1 Supervsed Learnng regardng support vector machnes (SVMs). Ths document also provdes a general
More information2016 Wiley. Study Session 2: Ethical and Professional Standards Application
6 Wley Study Sesson : Ethcal and Professonal Standards Applcaton LESSON : CORRECTION ANALYSIS Readng 9: Correlaton and Regresson LOS 9a: Calculate and nterpret a sample covarance and a sample correlaton
More informationUsing the estimated penetrances to determine the range of the underlying genetic model in casecontrol
Georgetown Unversty From the SelectedWorks of Mark J Meyer 8 Usng the estmated penetrances to determne the range of the underlyng genetc model n casecontrol desgn Mark J Meyer Neal Jeffres Gang Zheng Avalable
More informationStatistics for Business and Economics
Statstcs for Busness and Economcs Chapter 11 Smple Regresson Copyrght 010 Pearson Educaton, Inc. Publshng as Prentce Hall Ch. 11-1 11.1 Overvew of Lnear Models n An equaton can be ft to show the best lnear
More informationPopulation Design in Nonlinear Mixed Effects Multiple Response Models: extension of PFIM and evaluation by simulation with NONMEM and MONOLIX
Populaton Desgn n Nonlnear Mxed Effects Multple Response Models: extenson of PFIM and evaluaton by smulaton wth NONMEM and MONOLIX May 4th 007 Carolne Bazzol, Sylve Retout, France Mentré Inserm U738 Unversty
More informationStatistics II Final Exam 26/6/18
Statstcs II Fnal Exam 26/6/18 Academc Year 2017/18 Solutons Exam duraton: 2 h 30 mn 1. (3 ponts) A town hall s conductng a study to determne the amount of leftover food produced by the restaurants n the
More informationMarkov Chain Monte Carlo (MCMC), Gibbs Sampling, Metropolis Algorithms, and Simulated Annealing Bioinformatics Course Supplement
Markov Chan Monte Carlo MCMC, Gbbs Samplng, Metropols Algorthms, and Smulated Annealng 2001 Bonformatcs Course Supplement SNU Bontellgence Lab http://bsnuackr/ Outlne! Markov Chan Monte Carlo MCMC! Metropols-Hastngs
More informationDETERMINATION OF UNCERTAINTY ASSOCIATED WITH QUANTIZATION ERRORS USING THE BAYESIAN APPROACH
Proceedngs, XVII IMEKO World Congress, June 7, 3, Dubrovn, Croata Proceedngs, XVII IMEKO World Congress, June 7, 3, Dubrovn, Croata TC XVII IMEKO World Congress Metrology n the 3rd Mllennum June 7, 3,
More informationThe Multiple Classical Linear Regression Model (CLRM): Specification and Assumptions. 1. Introduction
ECONOMICS 5* -- NOTE (Summary) ECON 5* -- NOTE The Multple Classcal Lnear Regresson Model (CLRM): Specfcaton and Assumptons. Introducton CLRM stands for the Classcal Lnear Regresson Model. The CLRM s also
More informationSmall Area Interval Estimation
.. Small Area Interval Estmaton Partha Lahr Jont Program n Survey Methodology Unversty of Maryland, College Park (Based on jont work wth Masayo Yoshmor, Former JPSM Vstng PhD Student and Research Fellow
More informationThis column is a continuation of our previous column
Comparson of Goodness of Ft Statstcs for Lnear Regresson, Part II The authors contnue ther dscusson of the correlaton coeffcent n developng a calbraton for quanttatve analyss. Jerome Workman Jr. and Howard
More informationA New Method for Estimating Overdispersion. David Fletcher and Peter Green Department of Mathematics and Statistics
A New Method for Estmatng Overdsperson Davd Fletcher and Peter Green Department of Mathematcs and Statstcs Byron Morgan Insttute of Mathematcs, Statstcs and Actuaral Scence Unversty of Kent, England Overvew
More information4 Analysis of Variance (ANOVA) 5 ANOVA. 5.1 Introduction. 5.2 Fixed Effects ANOVA
4 Analyss of Varance (ANOVA) 5 ANOVA 51 Introducton ANOVA ANOVA s a way to estmate and test the means of multple populatons We wll start wth one-way ANOVA If the populatons ncluded n the study are selected
More informationBasic Business Statistics, 10/e
Chapter 13 13-1 Basc Busness Statstcs 11 th Edton Chapter 13 Smple Lnear Regresson Basc Busness Statstcs, 11e 009 Prentce-Hall, Inc. Chap 13-1 Learnng Objectves In ths chapter, you learn: How to use regresson
More informationBIO Lab 2: TWO-LEVEL NORMAL MODELS with school children popularity data
Lab : TWO-LEVEL NORMAL MODELS wth school chldren popularty data Purpose: Introduce basc two-level models for normally dstrbuted responses usng STATA. In partcular, we dscuss Random ntercept models wthout
More informationTesting for seasonal unit roots in heterogeneous panels
Testng for seasonal unt roots n heterogeneous panels Jesus Otero * Facultad de Economía Unversdad del Rosaro, Colomba Jeremy Smth Department of Economcs Unversty of arwck Monca Gulett Aston Busness School
More informationCokriging Partial Grades - Application to Block Modeling of Copper Deposits
Cokrgng Partal Grades - Applcaton to Block Modelng of Copper Deposts Serge Séguret 1, Julo Benscell 2 and Pablo Carrasco 2 Abstract Ths work concerns mneral deposts made of geologcal bodes such as breccas
More information4.3 Poisson Regression
of teratvely reweghted least squares regressons (the IRLS algorthm). We do wthout gvng further detals, but nstead focus on the practcal applcaton. > glm(survval~log(weght)+age, famly="bnomal", data=baby)
More informationSIMPLE LINEAR REGRESSION
Smple Lnear Regresson and Correlaton Introducton Prevousl, our attenton has been focused on one varable whch we desgnated b x. Frequentl, t s desrable to learn somethng about the relatonshp between two
More informationBayesian Planning of Hit-Miss Inspection Tests
Bayesan Plannng of Ht-Mss Inspecton Tests Yew-Meng Koh a and Wllam Q Meeker a a Center for Nondestructve Evaluaton, Department of Statstcs, Iowa State Unversty, Ames, Iowa 5000 Abstract Although some useful
More informationChapter 12 Analysis of Covariance
Chapter Analyss of Covarance Any scentfc experment s performed to know somethng that s unknown about a group of treatments and to test certan hypothess about the correspondng treatment effect When varablty
More informationx i1 =1 for all i (the constant ).
Chapter 5 The Multple Regresson Model Consder an economc model where the dependent varable s a functon of K explanatory varables. The economc model has the form: y = f ( x,x,..., ) xk Approxmate ths by
More informationExplaining the Stein Paradox
Explanng the Sten Paradox Kwong Hu Yung 1999/06/10 Abstract Ths report offers several ratonale for the Sten paradox. Sectons 1 and defnes the multvarate normal mean estmaton problem and ntroduces Sten
More informationTurbulence classification of load data by the frequency and severity of wind gusts. Oscar Moñux, DEWI GmbH Kevin Bleibler, DEWI GmbH
Turbulence classfcaton of load data by the frequency and severty of wnd gusts Introducton Oscar Moñux, DEWI GmbH Kevn Blebler, DEWI GmbH Durng the wnd turbne developng process, one of the most mportant
More informationChapter 8 Indicator Variables
Chapter 8 Indcator Varables In general, e explanatory varables n any regresson analyss are assumed to be quanttatve n nature. For example, e varables lke temperature, dstance, age etc. are quanttatve n
More informationCorrelation and Regression. Correlation 9.1. Correlation. Chapter 9
Chapter 9 Correlaton and Regresson 9. Correlaton Correlaton A correlaton s a relatonshp between two varables. The data can be represented b the ordered pars (, ) where s the ndependent (or eplanator) varable,
More informationIntroduction to Generalized Linear Models
INTRODUCTION TO STATISTICAL MODELLING TRINITY 00 Introducton to Generalzed Lnear Models I. Motvaton In ths lecture we extend the deas of lnear regresson to the more general dea of a generalzed lnear model
More informationLecture 16 Statistical Analysis in Biomaterials Research (Part II)
3.051J/0.340J 1 Lecture 16 Statstcal Analyss n Bomaterals Research (Part II) C. F Dstrbuton Allows comparson of varablty of behavor between populatons usng test of hypothess: σ x = σ x amed for Brtsh statstcan
More informationOpen Systems: Chemical Potential and Partial Molar Quantities Chemical Potential
Open Systems: Chemcal Potental and Partal Molar Quanttes Chemcal Potental For closed systems, we have derved the followng relatonshps: du = TdS pdv dh = TdS + Vdp da = SdT pdv dg = VdP SdT For open systems,
More informationis the calculated value of the dependent variable at point i. The best parameters have values that minimize the squares of the errors
Multple Lnear and Polynomal Regresson wth Statstcal Analyss Gven a set of data of measured (or observed) values of a dependent varable: y versus n ndependent varables x 1, x, x n, multple lnear regresson
More informationLinear Approximation with Regularization and Moving Least Squares
Lnear Approxmaton wth Regularzaton and Movng Least Squares Igor Grešovn May 007 Revson 4.6 (Revson : March 004). 5 4 3 0.5 3 3.5 4 Contents: Lnear Fttng...4. Weghted Least Squares n Functon Approxmaton...
More informationUsing T.O.M to Estimate Parameter of distributions that have not Single Exponential Family
IOSR Journal of Mathematcs IOSR-JM) ISSN: 2278-5728. Volume 3, Issue 3 Sep-Oct. 202), PP 44-48 www.osrjournals.org Usng T.O.M to Estmate Parameter of dstrbutons that have not Sngle Exponental Famly Jubran
More informationA Bayes Algorithm for the Multitask Pattern Recognition Problem Direct Approach
A Bayes Algorthm for the Multtask Pattern Recognton Problem Drect Approach Edward Puchala Wroclaw Unversty of Technology, Char of Systems and Computer etworks, Wybrzeze Wyspanskego 7, 50-370 Wroclaw, Poland
More informationPredictive Analytics : QM901.1x Prof U Dinesh Kumar, IIMB. All Rights Reserved, Indian Institute of Management Bangalore
Sesson Outlne Introducton to classfcaton problems and dscrete choce models. Introducton to Logstcs Regresson. Logstc functon and Logt functon. Maxmum Lkelhood Estmator (MLE) for estmaton of LR parameters.
More informationLOW BIAS INTEGRATED PATH ESTIMATORS. James M. Calvin
Proceedngs of the 007 Wnter Smulaton Conference S G Henderson, B Bller, M-H Hseh, J Shortle, J D Tew, and R R Barton, eds LOW BIAS INTEGRATED PATH ESTIMATORS James M Calvn Department of Computer Scence
More informationInterval Estimation in the Classical Normal Linear Regression Model. 1. Introduction
ECONOMICS 35* -- NOTE 7 ECON 35* -- NOTE 7 Interval Estmaton n the Classcal Normal Lnear Regresson Model Ths note outlnes the basc elements of nterval estmaton n the Classcal Normal Lnear Regresson Model
More information3.1 Expectation of Functions of Several Random Variables. )' be a k-dimensional discrete or continuous random vector, with joint PMF p (, E X E X1 E X
Statstcs 1: Probablty Theory II 37 3 EPECTATION OF SEVERAL RANDOM VARIABLES As n Probablty Theory I, the nterest n most stuatons les not on the actual dstrbuton of a random vector, but rather on a number
More informationRESIDUALS AND INFLUENCE IN NONLINEAR REGRESSION FOR REPEATED MEASUREMENT DATA
Operatons Research and Applcatons : An Internatonal Journal (ORAJ), Vol.4, No.3/4, November 17 RESIDUALS AND INFLUENCE IN NONLINEAR REGRESSION FOR REPEAED MEASUREMEN DAA Munsr Al, Yu Feng, Al choo, Zamr
More informationANOMALIES OF THE MAGNITUDE OF THE BIAS OF THE MAXIMUM LIKELIHOOD ESTIMATOR OF THE REGRESSION SLOPE
P a g e ANOMALIES OF THE MAGNITUDE OF THE BIAS OF THE MAXIMUM LIKELIHOOD ESTIMATOR OF THE REGRESSION SLOPE Darmud O Drscoll ¹, Donald E. Ramrez ² ¹ Head of Department of Mathematcs and Computer Studes
More informationModule 3 LOSSY IMAGE COMPRESSION SYSTEMS. Version 2 ECE IIT, Kharagpur
Module 3 LOSSY IMAGE COMPRESSION SYSTEMS Verson ECE IIT, Kharagpur Lesson 6 Theory of Quantzaton Verson ECE IIT, Kharagpur Instructonal Objectves At the end of ths lesson, the students should be able to:
More informationUSE OF DOUBLE SAMPLING SCHEME IN ESTIMATING THE MEAN OF STRATIFIED POPULATION UNDER NON-RESPONSE
STATISTICA, anno LXXV, n. 4, 015 USE OF DOUBLE SAMPLING SCHEME IN ESTIMATING THE MEAN OF STRATIFIED POPULATION UNDER NON-RESPONSE Manoj K. Chaudhary 1 Department of Statstcs, Banaras Hndu Unversty, Varanas,
More informationA Comparative Study for Estimation Parameters in Panel Data Model
A Comparatve Study for Estmaton Parameters n Panel Data Model Ahmed H. Youssef and Mohamed R. Abonazel hs paper examnes the panel data models when the regresson coeffcents are fxed random and mxed and
More informationChapter 3 Describing Data Using Numerical Measures
Chapter 3 Student Lecture Notes 3-1 Chapter 3 Descrbng Data Usng Numercal Measures Fall 2006 Fundamentals of Busness Statstcs 1 Chapter Goals To establsh the usefulness of summary measures of data. The
More informationRELIABILITY ASSESSMENT
CHAPTER Rsk Analyss n Engneerng and Economcs RELIABILITY ASSESSMENT A. J. Clark School of Engneerng Department of Cvl and Envronmental Engneerng 4a CHAPMAN HALL/CRC Rsk Analyss for Engneerng Department
More informationBAYESIAN CURVE FITTING USING PIECEWISE POLYNOMIALS. Dariusz Biskup
BAYESIAN CURVE FITTING USING PIECEWISE POLYNOMIALS Darusz Bskup 1. Introducton The paper presents a nonparaetrc procedure for estaton of an unknown functon f n the regresson odel y = f x + ε = N. (1) (
More informationSTAT 511 FINAL EXAM NAME Spring 2001
STAT 5 FINAL EXAM NAME Sprng Instructons: Ths s a closed book exam. No notes or books are allowed. ou may use a calculator but you are not allowed to store notes or formulas n the calculator. Please wrte
More informationLaboratory 3: Method of Least Squares
Laboratory 3: Method of Least Squares Introducton Consder the graph of expermental data n Fgure 1. In ths experment x s the ndependent varable and y the dependent varable. Clearly they are correlated wth
More informationLaboratory 1c: Method of Least Squares
Lab 1c, Least Squares Laboratory 1c: Method of Least Squares Introducton Consder the graph of expermental data n Fgure 1. In ths experment x s the ndependent varable and y the dependent varable. Clearly
More informationEffective plots to assess bias and precision in method comparison studies
Effectve plots to assess bas and precson n method comparson studes Bern, November, 016 Patrck Taffé, PhD Insttute of Socal and Preventve Medcne () Unversty of Lausanne, Swtzerland Patrck.Taffe@chuv.ch
More informationLecture Notes on Linear Regression
Lecture Notes on Lnear Regresson Feng L fl@sdueducn Shandong Unversty, Chna Lnear Regresson Problem In regresson problem, we am at predct a contnuous target value gven an nput feature vector We assume
More informationECONOMICS 351*-A Mid-Term Exam -- Fall Term 2000 Page 1 of 13 pages. QUEEN'S UNIVERSITY AT KINGSTON Department of Economics
ECOOMICS 35*-A Md-Term Exam -- Fall Term 000 Page of 3 pages QUEE'S UIVERSITY AT KIGSTO Department of Economcs ECOOMICS 35* - Secton A Introductory Econometrcs Fall Term 000 MID-TERM EAM ASWERS MG Abbott
More informationEstimation: Part 2. Chapter GREG estimation
Chapter 9 Estmaton: Part 2 9. GREG estmaton In Chapter 8, we have seen that the regresson estmator s an effcent estmator when there s a lnear relatonshp between y and x. In ths chapter, we generalzed the
More informationPolynomial Regression Models
LINEAR REGRESSION ANALYSIS MODULE XII Lecture - 6 Polynomal Regresson Models Dr. Shalabh Department of Mathematcs and Statstcs Indan Insttute of Technology Kanpur Test of sgnfcance To test the sgnfcance
More informationChapter 15 - Multiple Regression
Chapter - Multple Regresson Chapter - Multple Regresson Multple Regresson Model The equaton that descrbes how the dependent varable y s related to the ndependent varables x, x,... x p and an error term
More informationLinear Regression Analysis: Terminology and Notation
ECON 35* -- Secton : Basc Concepts of Regresson Analyss (Page ) Lnear Regresson Analyss: Termnology and Notaton Consder the generc verson of the smple (two-varable) lnear regresson model. It s represented
More informationChapter 5. Solution of System of Linear Equations. Module No. 6. Solution of Inconsistent and Ill Conditioned Systems
Numercal Analyss by Dr. Anta Pal Assstant Professor Department of Mathematcs Natonal Insttute of Technology Durgapur Durgapur-713209 emal: anta.bue@gmal.com 1 . Chapter 5 Soluton of System of Lnear Equatons
More informationSee Book Chapter 11 2 nd Edition (Chapter 10 1 st Edition)
Count Data Models See Book Chapter 11 2 nd Edton (Chapter 10 1 st Edton) Count data consst of non-negatve nteger values Examples: number of drver route changes per week, the number of trp departure changes
More informationChapter 2 - The Simple Linear Regression Model S =0. e i is a random error. S β2 β. This is a minimization problem. Solution is a calculus exercise.
Chapter - The Smple Lnear Regresson Model The lnear regresson equaton s: where y + = β + β e for =,..., y and are observable varables e s a random error How can an estmaton rule be constructed for the
More informationOn mutual information estimation for mixed-pair random variables
On mutual nformaton estmaton for mxed-par random varables November 3, 218 Aleksandr Beknazaryan, Xn Dang and Haln Sang 1 Department of Mathematcs, The Unversty of Msssspp, Unversty, MS 38677, USA. E-mal:
More informationInternational Journal of Engineering Research and Modern Education (IJERME) Impact Factor: 7.018, ISSN (Online): (
CONSTRUCTION AND SELECTION OF CHAIN SAMPLING PLAN WITH ZERO INFLATED POISSON DISTRIBUTION A. Palansamy* & M. Latha** * Research Scholar, Department of Statstcs, Government Arts College, Udumalpet, Tamlnadu
More informationANOVA. The Observations y ij
ANOVA Stands for ANalyss Of VArance But t s a test of dfferences n means The dea: The Observatons y j Treatment group = 1 = 2 = k y 11 y 21 y k,1 y 12 y 22 y k,2 y 1, n1 y 2, n2 y k, nk means: m 1 m 2
More information