Finite Mixtures of Multivariate Skew Laplace Distributions
|
|
- Russell Newman
- 5 years ago
- Views:
Transcription
1 Fiite Mixtures of Multivariate Skew Laplace Distributios Fatma Zehra Doğru 1 *, Y. Murat Bulut 2 ad Olcay Arsla 3 1 Giresu Uiversity, Faculty of Arts ad Scieces, Departmet of Statistics, Giresu/Turkey. ( fatma.dogru@giresu.edu.tr) 2 Eskisehir Osmagazi Uiversity, Faculty of Sciece ad Letters, Departmet of Statistics, Eskisehir/Turkey.( ymbulut@ogu.edu.tr) 3 Akara Uiversity, Faculty of Sciece, Departmet of Statistics, Akara/Turkey. ( oarsla@akara.edu.tr) Abstract I this paper, we propose fiite mixtures of multivariate skew Laplace distributios to model both skewess ad heavy-tailedess i the heterogeeous data sets. The maximum likelihood estimators for the parameters of iterest are obtaied by usig the EM algorithm. We give a small simulatio study ad a real data example to illustrate the performace of the proposed mixture model. Key words: EM algorithm, ML estimatio, multivariate mixture model, MSL. 1. Itroductio Fiite mixture models are used for modelig heterogeeous data sets as a result of their flexibility. These models are commoly applied i fields such as classificatio, cluster ad latet class aalysis, desity estimatio, data miig, image aalysis, geetics, medicie, patter recogitio etc. (see for more detail, Titterigto et al. (1985), McLachla ad Basford (1988), McLachla ad Peel (2000), Bishop (2006), Frühwirth-Schatter (2006)). Geerally, due to its tractability ad wide applicability, the distributio of mixture model compoets is assumed to be ormal. However, i practice, the data sets may be asymmetric ad/or heavy-tailed. For this, there are some studies i literature for multivariate mixture modelig usig the asymmetric ad/or heavy-tailed distributios. Some of these studies ca be summarized as follows. Peel ad McLachla (2000) proposed fiite mixtures of multivariate t distributios as a robust extesio of the multivariate ormal mixture model (McLachla ad Basford (1988)), Li (2009) itroduced multivariate skew ormal mixture models, Pye et al. (2009) ad Li (2010) proposed fiite mixtures of the restricted ad urestricted variats of multivariate skew t distributios of Sahu et al. (2003), Cabral et al. (2012) proposed multivariate mixture modelig based o the skew-ormal idepedet distributios ad Li et al. (2014) itroduced a flexible mixture modelig based o the skew-t-ormal distributio. I the multivariate aalysis, the multivariate skew ormal (MSN) (Azzalii ad Dalla Valle (1996), Gupta et al. (2004) ad Arellao-Valle ad Geto (2005)) distributio has bee proposed as a alterative to the multivariate ormal (MN) distributio to deal with skewess i the data. However, sice MSN distributio is ot heavy-tailed some alterative heavy-tailed skew distributios are eeded to model skewess ad heavy-tailedess. Oe of the examples of heavy-tailed skew distributio is the multivariate skew t (MST) distributio defied by Azzalii ad Capitaio (2003) ad Gupta (2003). The other heavy-tailed skew distributio is the multivariate skew Laplace (MSL) distributio proposed by Arsla (2010). The advatage of this distributio is that the MSL distributio has the less umber of parameters tha MST distributio ad it has the same umber of parameters with the MSN distributio. Cocerig the fiite mixtures of distributios, sice the mixtures of MN distributios are ot able to 1
2 model skewess, fiite mixtures of MSN distributios have bee proposed by Li (2009) to model such data sets. I this study, we explore the fiite mixtures of MSL distributios as a alterative to the fiite mixtures of MSN distributios to deal with both skewess ad heavy-tailedess i the heterogeeous data sets. The rest of this paper is orgaized as follows. I Sectio 2, we briefly summarize some properties of MSL distributio. I Sectio 3, we preset the mixtures of MSL distributios ad give the Expectatio- Maximizatio (EM) algorithm to obtai the maximum likelihood (ML) estimators of proposed mixture model. I Sectio 4, we give the empirical iformatio matrix of MSL distributio to compute the stadard errors of proposed estimators. I the Applicatio Sectio, we give a small simulatio study ad a real data example to illustrate the performace of proposed mixture model. Some coclusios are give i Sectio Multivariate skew Laplace distributio A p-dimesioal radom vector Y R p is said to have MSL distributio (Y MSL p (μ, Σ, γ)) which is give by Arsla (2010) if it has the followig probability desity fuctio (pdf) f MSL (y; μ, Σ, γ) = Σ p π p 1 2 αγ ( p ) exp { α (y μ)t Σ 1 (y μ) + (y μ)t Σ 1 γ}, (1) where α = 1 + γ T Σ 1 γ, μ R p is a locatio parameter, γ R p is a skewess parameter ad Σ is a positive defiite scatter matrix. Propositio 1. The characteristic fuctio of MSL (μ, Σ, γ) is Φ Y (t) = e ittμ [1 + t T Σt 2it T γ] (p+1)/2, tεr p. Proof. Sice we kow coditioal distributio of Y give V, we get characteristic fuctio as follows Φ Y (t) = E (E Y V (e itty )) = e ittμ E V (e V 1 ( tt Σt 2 itt γ) ) = e ittμ e v 1 ( tt Σt 2 itt γ) g(v)dv 0 = e ittμ [1 + t T Σt 2it T γ] (+1)/2. If Y MSL p (μ, Σ, γ) the the expectatio ad variace of Y is E(Y) = μ + (p + 1)γ, Var(Y) = (p + 1)(Σ + 2γγ T ). The MSL distributio ca be obtaied as a variace-mea mixture of MN distributio ad iverse gamma (IG) distributio. The variace-mea mixture represetatio is give as follows Y = μ + V 1 γ + V 1 Σ 1/2 X (2) 2
3 where X N p (0, I p ) ad V IG( p+1, 1 ). Note that if γ = 0, the desity fuctio of Y reduces to the 2 2 desity fuctio of symmetric multivariate Laplace distributio give by Naik ad Plugpogpu (2006). Also, the coditioal distributio of Y give V = v will be Y v N(μ + v 1 γ, v 1 Σ). The joit desity fuctio of Y ad V is f(y, v) = 1 Σ 2e (y μ)t Σ 1 γ 2 p π p 1 2 αγ ( p + 1 {v 3 2e 1 2 {(y μ)t Σ 1 (y μ)v+(1+γ T Σ 1 γ)v 1} }. 2 ) The, we have the followig coditioal desity fuctio of V give Y f(v y) = α 2π eα (y μ)t Σ 1 (y μ) v 3 2 e 1 2 {(y μ)t Σ 1 (y μ)v+α 2 v 1}, v > 0. (3) Propositio 2. Usig the coditioal desity fuctio give i (3), the coditioal expectatios ca be obtaied as follows 1 + γ T Σ 1 γ E(V y) = (y μ) T Σ 1 (y μ), (4) E(V 1 y) = 1 + (1 + γt Σ 1 γ)(y μ) T Σ 1 (y μ) 1 + γ T Σ 1. γ (5) Note that these coditioal expectatios will be used i the EM algorithm give i Subsectio Fiite mixtures of the MSL distributios Let y 1,, y be a p-dimesioal radom sample which come from a g-compoet mixtures of MSL distributios. The pdf of a g-compoet fiite mixtures of MSL distributios is give by g f(y Θ) = π i f(y; μ i, Σ i, γ i ), i=1 (6) g where π i deotes the mixig probability with i=1 π i = 1, 0 π i 1, f(y; μ i, Σ i, λ i ) represets the pdf of the ith compoet (pdf of the MSL distributio) give i (1) ad Θ = (π 1,, π g, μ 1,, μ g, Σ 1,, Σ g, γ 1,, γ g ) T is the ukow parameter vector. 3.1 ML estimatio The ML estimator of Θ ca be foud by maximizig the followig log-likelihood fuctio g l(θ) = log ( π i f(y j ; μ i, Σ i, γ i )). i=1 (7) 3
4 However, there is ot a explicit maximizer of (7). Therefore, i geeral, the EM algorithm (Dempster et al. (1977)) is used to obtai the ML estimator of Θ. Here, we will use the followig EM algorithm. Let z j = (z 1j,, z gj ) T be the latet variables with z ij = { 1, if jth observatio belogs to i th compoet 0, otherwise (8) where j = 1,, ad i = 1,, g. To implemet the steps of the EM algorithm, we will use the stochastic represetatio of the MSL distributio give i (2). The, the hierarchical represetatio for the mixtures of MSL distributios will be Y j v j, z ij = 1 N(μ + v j 1 γ, v j 1 Σ), V j z ij = 1 IG ( p + 1 2, 1 2 ). (9) Let (y, v, z) be the complete data, where y = (y 1 T,, y T ) T, v = (v 1,, v ) ad z = (z 1,, z ) T. Usig the hierarchical represetatio give above ad igorig the costats, the complete data loglikelihood fuctio ca be writte by l c (Θ; y, v, z) = z ij {log π i 1 2 log Σ i + (y j μ i ) T Σ 1 i γ i i=1 1 2 v j(y j μ i ) T Σ i 1 (y j μ i ) 1 2 γ i T Σ i 1 γ i v j (3 log v j + v j 1 )}. (10) To overcome the latecy of the latet variables give i (10), we have to take the coditioal expectatio of the complete data log-likelihood fuctio give the observed data y j g E(l c (Θ; y, v, z) y j ) = E(z ij y j ) {log π i 1 2 log Σ i (y j μ i ) T Σ 1 i γ i i=1 1 2 E(V j y j )(y j μ i ) T Σ i 1 (y j μ i ) 1 2 γ i T Σ i 1 γ i E(V j 1 y j )}. (11) Sice the last part of the complete data log-likelihood fuctio does ot iclude the parameters, the above coditioal expectatio of the complete data log-likelihood fuctio oly cosists of the ecessary coditioal expectatios. Note that the coditioal expectatios E(V j y j ) ad E(V j 1 y j ) ca be calculated usig the coditioal expectatios give i (4) ad (5) ad the coditioal expectatio E(z ij y j ) ca be computed usig the classical theory of mixture modelig. EM algorithm: 1. Set iitial parameter estimate Θ ad a stoppig rule Δ. 2. E-Step: Compute the followig coditioal expectatios for k = 0,1,2, iteratio = E(zij y j, Θ ) = π i f (yj ; μ i, Σ i, γ i ) f(y j ; Θ ), (12) 4
5 v 1ij = E(Vj y j, Θ ) = 1 + T 1 γ i Σ i γ i, T (y j 1 μ i ) Σ i (y j μ i ) (13) v 2ij = 1 E(Vj y j, Θ ) = 1 + (1 + T 1 T γ i Σ i γ i ) (yj 1 μ i ) Σ i (y j μ i ) 1 + T 1 γ i Σ i γ i. (14) The, we form the followig objective fuctio Q(Θ; Θ ) = g i=1 {log π i 1 2 log Σ i (y j μ i ) T Σ i 1 γ i 1 2 v 1ij (yj μ i ) T Σ 1 i (y j μ i ) 1 2 v 2ij T γi Σ 1 i γ i }. (15) 3. M-step: Maximize the Q(Θ; Θ ) with respect to Θ to get the (k + 1)th parameter estimates for the parameters. This maximizatio yields the followig updatig equatios (k+1) π i = (k+1) μ i = v 1ij, v 1ij yj v 1ij γ i, ( (k+1) ) ( y j ) ( ) ( ) γ i = ( v 1ij ) ( v 2ij ) ( ) 2, (18) (k+1) Σ i = T v 1ij (yj μ i ) (yj μ i ) γ i γ i T v 1ij yj v 2ij 4. Repeat E ad M steps util the covergece rule Θ (k+1) Θ < Δ is obtaied. Alteratively, the absolute differece of the actual log-likelihood l(θ (k+1) ) l(θ ) < Δ or l(θ (k+1) ) l(θ ) 1 < Δ ca be used as a stoppig rule (see Dias ad Wedel (2004)).. (16) (17) (19) 3.2 Iitial values To determie the iitial values for the EM algorithm, we will use the same procedure give i Li (2009). The steps of the selectig iitial values are give as follows. i) Perform the K-meas clusterig algorithm (Hartiga ad Wog (1979)). g ii) Iitialize the compoet labels z j = {zij } accordig to the K-meas clusterig results. i=1 iii) The iitial values of mixig probabilities, compoet locatios ad compoet scale variaces ca be set as 5
6 π i = Σ i = y j, μ i = T (y j μ i ) (yj μ i ) iv) For the skewess parameters, use the skewess coefficiet vector of each clusters.,. 4. The empirical iformatio matrix We will compute the stadard errors of ML estimators usig the iformatio based method give by Basford et al. (1997). Here, we will use the iverse of the empirical iformatio matrix to have a approximatio to the asymptotic covariace matrix of estimators. The iformatio matrix is I e = s js jt, (20) where s j = E Θ ( l cj(θ;y j,v j,z j ) y Θ j ), j = 1,, are the idividual scores ad l cj (Θ; y j, v j, z j ) is the complete data log-likelihood fuctio for the jth observatio. The, the elemets of the score vector s j is (s j,π1,, s j,πg 1, s j,μ1,, s j,μg, s j,σ1,, s j,σg, s j,γ1,, s j,γg ) T. After some straightforward algebra, we obtai the followig equatios s j,πr = z rj z gj π r π g, r = 1,, g 1, (21) s j,μi = Σ i 1 (v 1ij (y j μ i) γ i), (22) s j,σi = vech ( { (Σ i 1 v 1ij Σ i 1 (y j μ i)(y j μ i) Σ i 1 v 2ij Σ i 1 γ iγ it Σ i 1 ) diag (Σ i 1 v 1ij Σ i 1 (y j μ i)(y j μ i) Σ i 1 (23) v 2ij Σ i 1 γ iγ it )} ), s j,γi = Σ i 1 ((y j μ i) v 2ij γ i). (24) Thus, the stadard errors of Θ will be foud by usig the square root of the diagoal elemets of the iverse of (20). 5. Applicatios I this sectio, we will illustrate the performace of proposed mixture model with a small simulatio study ad a real data example. All computatios for simulatio study ad real data example were coducted usig MATLAB R2013a. For all computatios, the stoppig rule Δ is take as Simulatio study I the simulatio study, the data are geerated from the followig two-compoet mixtures of MSL distributios 6
7 f(y j Θ) = π 1 f p (y j ; μ 1, Σ 1, γ 1 ) + (1 π 1 )f p (y j ; μ 2, Σ 2, γ 2 ), where σ i,12 μ i = (μ i1, μ i2 ) T, Σ i = [ σ i,11 σ i,21 σ ], γ i = (γ i1, γ i2 ) T, i = 1,2 i,22 with the parameter values μ 1 = (2,2) T, μ 2 = ( 2, 2) T, Σ 1 = Σ 2 = [ ], γ 1 = (1,1) T, γ 2 = ( 1, 1) T, π 1 = 0.6. We set sample sizes as 500, 1000 ad We take the umber of replicates (N) as 500. The tables cotai the mea ad mea Euclidea distace values of estimates. For istace, the formula for mea Euclidea distace of μ i is give below μ i μ i = 1 N N ((μ ij μ ij ) 2 ) 1 2. Similarly, the other mea Euclidea distaces for the other estimates are obtaied. Note that for the π 1 the distace will be mea squared error (MSE). The formula of MSE is give MSE (π ) = 1 N N (π j π) 2 where π is the true parameter value, π j is the estimate of π for the jth simulated data ad π = 1 N N. Table 1 shows the simulatio results for the sample sizes 500, 1000 ad I the tables, we give mea ad mea Euclidea distace values of estimates ad true parameter values. We ca observe from this table that the proposed model is workig accurately to obtai the estimates for all parameters. This ca be observed from mea Euclidia distaces that are gettig smaller whe the sample sizes icrease. π j 7
8 Table 1. Mea ad mea Euclidea distace values of estimates for = 500, 1000 ad Compoets 1 2 Parameter True Mea Distace True Mea Distace π μ i1 μ i σ i, σ i, σ i, γ i γ i π μ i μ i σ i, σ i, σ i, γ i γ i π μ i μ i σ i, σ i, σ i, γ i γ i Real data example I this real data example, we will ivestigate the bak data set which was give i Tables 1.1 ad 1.2 by Flury ad Riedwyl (1988) ad examied by Ma ad Geto (2004) to model with a skew-symmetric distributio. There are six measuremets made o 100 geuie ad 100 couterfeit old Swiss 1000 frac bills for this data set. This data set also aalyzed by Li (2009) to model mixtures of MSN distributios. They used the sample of X 1 : the width of the right edge ad X 2 : the legth of the image diagoal which reveals a bimodal distributio with asymmetric compoets. I this study, we will use the Swiss bak data to illustrate the applicability of the fiite mixtures of multivariate skew Laplace distributios (FM- MSL) ad also we compare the results with the fiite mixtures of multivariate skew ormal distributios (FM-MSN) which was give by Li (2009). We give the estimatio results i Table 2 for FM-MSN ad FM-MSL. The table cosists of the ML estimates, stadard errors of estimates for all compoets, the log-likelihood, the values of the Akaike iformatio criterio (AIC) (Akaike (1973)) ad the Bayesia iformatio criterio (BIC) (Schwarz (1978)) for the FM-MSL. Also, we give the estimatio results ad criterio values for FM-MSN which was computed by Li (2009) i Table 2. Accordig to iformatio criterio values that the FM-MSL has better fit tha the FM-MSN. I Figure 1, we display the scatter plot of the data alog with the cotour plots of the fitted two-compoet FM-MSL model. We ca observe from figure that the FM-MSL captures the asymmetry ad accurately fit the data. 8
9 Table 2. ML estimatio results of the Swiss bak data set for FM-MSN ad FM-MSL. FM-MSN FM-MSL Estimate SE Estimate SE Estimate SE Estimate SE w μ i μ i σ i, σ i, σ i, γ i γ i l(θ ) AIC BIC Figure 1. Scatter plot of the Swiss bak data alog with the cotour plots of the fitted twocompoet FM-MSL model 6. Coclusios I this paper, we have proposed the mixtures of MSL distributios. We have give the EM algorithm to obtai the estimates. We have provided a small simulatio study to show the performace of proposed mixture model. We have observed from simulatio study that the proposed mixture model has accurately estimated the parameters. We have also give a real data example to compare the mixtures of MSL distributios with the mixtures of MSN distributios. We have observed from this example that the proposed model has the best fit accordig to the iformatio criterio values. Thus, the proposed model ca be used as a alterative mixture model to the mixtures of MSN distributios. Refereces Akaike, H Iformatio theory ad a extesio of the maximum likelihood priciple. Proceedig of the Secod Iteratioal Symposium o Iformatio Theory, B.N. Petrov ad F. Caski, eds., , Akademiai Kiado, Budapest. 9
10 Arellao-Valle, R.B. ad Geto, M.G O fudametal skew distributios. Joural of Multivariate Aalysis, 96(1), Arsla, O A alterative multivariate skew Laplace distributio: properties ad estimatio. Statistical Papers, 51(4), Azzalii, A. ad Dalla Valle, A The multivariate skew-ormal distributio. Biometrika, 83(4), Azzalii, A. ad Capitaio, A Distributios geerated by perturbatio of symmetry with emphasis o a multivariate skew t distributio. Joural of the Royal Statistical Society: Series B (Statistical Methodology), 65(2), Basford, K.E., Greeway, D.R., McLachla, G.J. ad Peel, D Stadard errors of fitted meas uder ormal mixture. Computatioal Statistics, 12, Bishop, C.M Patter Recogitio ad Machie Learig. Spriger, Sigapore. Cabral, C.R.B., Lachos, V.H. ad Prates, M.O Multivariate mixture modelig usig skew-ormal idepedet distributios. Computatioal Statistics & Data Aalysis, 56(1), Dempster, A.P., Laird, N.M. ad Rubi, D.B Maximum likelihood from icomplete data via the EM algorithm. Joural of the Royal Statistical Society, Series B, 39, Dias, J.G. ad Wedel, M A empirical compariso of EM, SEM ad MCMC performace for problematic gaussia mixture likelihoods. Statistics ad Computig, 14, Flury, B. ad H. Riedwyl Multivariate Statistics, a Practical Approach, Cambridge Uiversity Press, Cambridge. Frühwirth-Schatter, S Fiite Mixture ad Markov Switchig Models. Spriger, New York. Gupta, A. K Multivariate skew t-distributio. Statistics: A Joural of Theoretical ad Applied Statistics, 37(4), Gupta, A.K., Gozález-Farıás, G. ad Domıńguez-Molia, J.A A multivariate skew ormal distributio. Joural of multivariate aalysis, 89(1), Hartiga, J. A. ad Wog, M. A Algorithm AS 136: A k-meas clusterig algorithm. Joural of the Royal Statistical Society. Series C (Applied Statistics), 28(1), Li, T.I Maximum likelihood estimatio for multivariate skew ormal mixture models. Joural of Multivariate Aalysis, 100, Li, T.I Robust mixture modelig usig multivariate skew t distributios. Statistics ad Computig, 20(3), Li, T.I., Ho, H. J. ad Lee, C.R Flexible mixture modellig usig the multivariate skew-t-ormal distributio. Statistics ad Computig, 24(4), Ma, Y. ad Geto, M.G Flexible class of skew-symmetric distribtios, Scadiavia Joural of Statististics, 31, McLachla, G.J. ad Basford, K.E Mixture Models: Iferece ad Applicatio to Clusterig. Marcel Dekker, New York. McLachla, G.J. ad Peel, D Fiite Mixture Models. Wiley, New York. Naik, D.N. ad Plugpogpu, K A Kotz-type distributio for multivariate statistical iferece. I Advaces i distributio theory, order statistics, ad iferece (pp ). Birkhäuser Bosto. Peel, D. ad McLachla, G. J Robust mixture modellig usig the t distributio. Statistics ad computig, 10(4), Pye, S., Hu, X., Wag, K., Rossi, E., Li, T.I., Maier, L., Baecher-Alla, C., McLachla, G.J., Tamayo, P., Hafler, D.A., De Jager, P.L. ad Mesirov, J.P Automated high-dimesioal flow cytometric data aalysis. Proc. Natl. Acad. Sci. USA 106, Sahu, S. K., Dey, D. K. ad Braco, M. D A ew class of multivariate skew distributios with applicatios to Bayesia regressio models. Caadia Joural of Statistics, 31(2), Schwarz, G Estimatig the dimesio of a model. The Aals of Statistics, 6(2), Titterigto, D.M., Smith, A.F.M. ad Markov, U.E Statistical Aalysis of Fiite Mixture Distributios. Wiley, New York. 10
Computing the maximum likelihood estimates: concentrated likelihood, EM-algorithm. Dmitry Pavlyuk
Computig the maximum likelihood estimates: cocetrated likelihood, EM-algorithm Dmitry Pavlyuk The Mathematical Semiar, Trasport ad Telecommuicatio Istitute, Riga, 13.05.2016 Presetatio outlie 1. Basics
More informationBayesian and E- Bayesian Method of Estimation of Parameter of Rayleigh Distribution- A Bayesian Approach under Linex Loss Function
Iteratioal Joural of Statistics ad Systems ISSN 973-2675 Volume 12, Number 4 (217), pp. 791-796 Research Idia Publicatios http://www.ripublicatio.com Bayesia ad E- Bayesia Method of Estimatio of Parameter
More informationA Note on Box-Cox Quantile Regression Estimation of the Parameters of the Generalized Pareto Distribution
A Note o Box-Cox Quatile Regressio Estimatio of the Parameters of the Geeralized Pareto Distributio JM va Zyl Abstract: Makig use of the quatile equatio, Box-Cox regressio ad Laplace distributed disturbaces,
More informationExpectation and Variance of a random variable
Chapter 11 Expectatio ad Variace of a radom variable The aim of this lecture is to defie ad itroduce mathematical Expectatio ad variace of a fuctio of discrete & cotiuous radom variables ad the distributio
More informationIntroducing a Novel Bivariate Generalized Skew-Symmetric Normal Distribution
Joural of mathematics ad computer Sciece 7 (03) 66-7 Article history: Received April 03 Accepted May 03 Available olie Jue 03 Itroducig a Novel Bivariate Geeralized Skew-Symmetric Normal Distributio Behrouz
More informationPAijpam.eu ON TENSOR PRODUCT DECOMPOSITION
Iteratioal Joural of Pure ad Applied Mathematics Volume 103 No 3 2015, 537-545 ISSN: 1311-8080 (prited versio); ISSN: 1314-3395 (o-lie versio) url: http://wwwijpameu doi: http://dxdoiorg/1012732/ijpamv103i314
More informationExpectation-Maximization Algorithm.
Expectatio-Maximizatio Algorithm. Petr Pošík Czech Techical Uiversity i Prague Faculty of Electrical Egieerig Dept. of Cyberetics MLE 2 Likelihood.........................................................................................................
More informationThe Expectation-Maximization (EM) Algorithm
The Expectatio-Maximizatio (EM) Algorithm Readig Assigmets T. Mitchell, Machie Learig, McGraw-Hill, 997 (sectio 6.2, hard copy). S. Gog et al. Dyamic Visio: From Images to Face Recogitio, Imperial College
More informationDistributional Similarity Models (cont.)
Distributioal Similarity Models (cot.) Regia Barzilay EECS Departmet MIT October 19, 2004 Sematic Similarity Vector Space Model Similarity Measures cosie Euclidea distace... Clusterig k-meas hierarchical
More informationR. van Zyl 1, A.J. van der Merwe 2. Quintiles International, University of the Free State
Bayesia Cotrol Charts for the Two-parameter Expoetial Distributio if the Locatio Parameter Ca Take o Ay Value Betwee Mius Iity ad Plus Iity R. va Zyl, A.J. va der Merwe 2 Quitiles Iteratioal, ruaavz@gmail.com
More informationDistributional Similarity Models (cont.)
Sematic Similarity Vector Space Model Similarity Measures cosie Euclidea distace... Clusterig k-meas hierarchical Last Time EM Clusterig Soft versio of K-meas clusterig Iput: m dimesioal objects X = {
More informationA Note on Effi cient Conditional Simulation of Gaussian Distributions. April 2010
A Note o Effi ciet Coditioal Simulatio of Gaussia Distributios A D D C S S, U B C, V, BC, C April 2010 A Cosider a multivariate Gaussia radom vector which ca be partitioed ito observed ad uobserved compoetswe
More informationThe new class of Kummer beta generalized distributions
The ew class of Kummer beta geeralized distributios Rodrigo Rossetto Pescim 12 Clarice Garcia Borges Demétrio 1 Gauss Moutiho Cordeiro 3 Saralees Nadarajah 4 Edwi Moisés Marcos Ortega 1 1 Itroductio Geeralized
More informationApproximate Confidence Interval for the Reciprocal of a Normal Mean with a Known Coefficient of Variation
Metodološki zvezki, Vol. 13, No., 016, 117-130 Approximate Cofidece Iterval for the Reciprocal of a Normal Mea with a Kow Coefficiet of Variatio Wararit Paichkitkosolkul 1 Abstract A approximate cofidece
More informationLecture 11 and 12: Basic estimation theory
Lecture ad 2: Basic estimatio theory Sprig 202 - EE 94 Networked estimatio ad cotrol Prof. Kha March 2 202 I. MAXIMUM-LIKELIHOOD ESTIMATORS The maximum likelihood priciple is deceptively simple. Louis
More informationGoodness-Of-Fit For The Generalized Exponential Distribution. Abstract
Goodess-Of-Fit For The Geeralized Expoetial Distributio By Amal S. Hassa stitute of Statistical Studies & Research Cairo Uiversity Abstract Recetly a ew distributio called geeralized expoetial or expoetiated
More informationANOTHER WEIGHTED WEIBULL DISTRIBUTION FROM AZZALINI S FAMILY
ANOTHER WEIGHTED WEIBULL DISTRIBUTION FROM AZZALINI S FAMILY Sulema Nasiru, MSc. Departmet of Statistics, Faculty of Mathematical Scieces, Uiversity for Developmet Studies, Navrogo, Upper East Regio, Ghaa,
More informationClustering. CM226: Machine Learning for Bioinformatics. Fall Sriram Sankararaman Acknowledgments: Fei Sha, Ameet Talwalkar.
Clusterig CM226: Machie Learig for Bioiformatics. Fall 216 Sriram Sakararama Ackowledgmets: Fei Sha, Ameet Talwalkar Clusterig 1 / 42 Admiistratio HW 1 due o Moday. Email/post o CCLE if you have questios.
More informationEstimating Confidence Interval of Mean Using. Classical, Bayesian, and Bootstrap Approaches
Iteratioal Joural of Mathematical Aalysis Vol. 8, 2014, o. 48, 2375-2383 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/10.12988/ijma.2014.49287 Estimatig Cofidece Iterval of Mea Usig Classical, Bayesia,
More informationPOWER AKASH DISTRIBUTION AND ITS APPLICATION
POWER AKASH DISTRIBUTION AND ITS APPLICATION Rama SHANKER PhD, Uiversity Professor, Departmet of Statistics, College of Sciece, Eritrea Istitute of Techology, Asmara, Eritrea E-mail: shakerrama009@gmail.com
More informationConfidence interval for the two-parameter exponentiated Gumbel distribution based on record values
Iteratioal Joural of Applied Operatioal Research Vol. 4 No. 1 pp. 61-68 Witer 2014 Joural homepage: www.ijorlu.ir Cofidece iterval for the two-parameter expoetiated Gumbel distributio based o record values
More informationA goodness-of-fit test based on the empirical characteristic function and a comparison of tests for normality
A goodess-of-fit test based o the empirical characteristic fuctio ad a compariso of tests for ormality J. Marti va Zyl Departmet of Mathematical Statistics ad Actuarial Sciece, Uiversity of the Free State,
More informationIt should be unbiased, or approximately unbiased. Variance of the variance estimator should be small. That is, the variance estimator is stable.
Chapter 10 Variace Estimatio 10.1 Itroductio Variace estimatio is a importat practical problem i survey samplig. Variace estimates are used i two purposes. Oe is the aalytic purpose such as costructig
More informationExponential Families and Bayesian Inference
Computer Visio Expoetial Families ad Bayesia Iferece Lecture Expoetial Families A expoetial family of distributios is a d-parameter family f(x; havig the followig form: f(x; = h(xe g(t T (x B(, (. where
More informationAlgorithms for Clustering
CR2: Statistical Learig & Applicatios Algorithms for Clusterig Lecturer: J. Salmo Scribe: A. Alcolei Settig: give a data set X R p where is the umber of observatio ad p is the umber of features, we wat
More informationThe Sampling Distribution of the Maximum. Likelihood Estimators for the Parameters of. Beta-Binomial Distribution
Iteratioal Mathematical Forum, Vol. 8, 2013, o. 26, 1263-1277 HIKARI Ltd, www.m-hikari.com http://d.doi.org/10.12988/imf.2013.3475 The Samplig Distributio of the Maimum Likelihood Estimators for the Parameters
More informationInvestigating the Significance of a Correlation Coefficient using Jackknife Estimates
Iteratioal Joural of Scieces: Basic ad Applied Research (IJSBAR) ISSN 2307-4531 (Prit & Olie) http://gssrr.org/idex.php?joural=jouralofbasicadapplied ---------------------------------------------------------------------------------------------------------------------------
More informationG. R. Pasha Department of Statistics Bahauddin Zakariya University Multan, Pakistan
Deviatio of the Variaces of Classical Estimators ad Negative Iteger Momet Estimator from Miimum Variace Boud with Referece to Maxwell Distributio G. R. Pasha Departmet of Statistics Bahauddi Zakariya Uiversity
More informationA proposed discrete distribution for the statistical modeling of
It. Statistical Ist.: Proc. 58th World Statistical Cogress, 0, Dubli (Sessio CPS047) p.5059 A proposed discrete distributio for the statistical modelig of Likert data Kidd, Marti Cetre for Statistical
More informationCEE 522 Autumn Uncertainty Concepts for Geotechnical Engineering
CEE 5 Autum 005 Ucertaity Cocepts for Geotechical Egieerig Basic Termiology Set A set is a collectio of (mutually exclusive) objects or evets. The sample space is the (collectively exhaustive) collectio
More informationTopic 9: Sampling Distributions of Estimators
Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be
More informationMOMENT-METHOD ESTIMATION BASED ON CENSORED SAMPLE
Vol. 8 o. Joural of Systems Sciece ad Complexity Apr., 5 MOMET-METHOD ESTIMATIO BASED O CESORED SAMPLE I Zhogxi Departmet of Mathematics, East Chia Uiversity of Sciece ad Techology, Shaghai 37, Chia. Email:
More informationTopic 9: Sampling Distributions of Estimators
Topic 9: Samplig Distributios of Estimators Course 003, 2018 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be
More informationA Risk Comparison of Ordinary Least Squares vs Ridge Regression
Joural of Machie Learig Research 14 (2013) 1505-1511 Submitted 5/12; Revised 3/13; Published 6/13 A Risk Compariso of Ordiary Least Squares vs Ridge Regressio Paramveer S. Dhillo Departmet of Computer
More information4.5 Multiple Imputation
45 ultiple Imputatio Itroductio Assume a parametric model: y fy x; θ We are iterested i makig iferece about θ I Bayesia approach, we wat to make iferece about θ from fθ x, y = πθfy x, θ πθfy x, θdθ where
More informationDouble Stage Shrinkage Estimator of Two Parameters. Generalized Exponential Distribution
Iteratioal Mathematical Forum, Vol., 3, o. 3, 3-53 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/.9/imf.3.335 Double Stage Shrikage Estimator of Two Parameters Geeralized Expoetial Distributio Alaa M.
More informationChapter 12 EM algorithms The Expectation-Maximization (EM) algorithm is a maximum likelihood method for models that have hidden variables eg. Gaussian
Chapter 2 EM algorithms The Expectatio-Maximizatio (EM) algorithm is a maximum likelihood method for models that have hidde variables eg. Gaussia Mixture Models (GMMs), Liear Dyamic Systems (LDSs) ad Hidde
More informationAccess to the published version may require journal subscription. Published with permission from: Elsevier.
This is a author produced versio of a paper published i Statistics ad Probability Letters. This paper has bee peer-reviewed, it does ot iclude the joural pagiatio. Citatio for the published paper: Forkma,
More informationKolmogorov-Smirnov type Tests for Local Gaussianity in High-Frequency Data
Proceedigs 59th ISI World Statistics Cogress, 5-30 August 013, Hog Kog (Sessio STS046) p.09 Kolmogorov-Smirov type Tests for Local Gaussiaity i High-Frequecy Data George Tauche, Duke Uiversity Viktor Todorov,
More informationPrecise Rates in Complete Moment Convergence for Negatively Associated Sequences
Commuicatios of the Korea Statistical Society 29, Vol. 16, No. 5, 841 849 Precise Rates i Complete Momet Covergece for Negatively Associated Sequeces Dae-Hee Ryu 1,a a Departmet of Computer Sciece, ChugWoo
More informationDiscrete Orthogonal Moment Features Using Chebyshev Polynomials
Discrete Orthogoal Momet Features Usig Chebyshev Polyomials R. Mukuda, 1 S.H.Og ad P.A. Lee 3 1 Faculty of Iformatio Sciece ad Techology, Multimedia Uiversity 75450 Malacca, Malaysia. Istitute of Mathematical
More informationEstimation of Gumbel Parameters under Ranked Set Sampling
Joural of Moder Applied Statistical Methods Volume 13 Issue 2 Article 11-2014 Estimatio of Gumbel Parameters uder Raked Set Samplig Omar M. Yousef Al Balqa' Applied Uiversity, Zarqa, Jorda, abuyaza_o@yahoo.com
More informationCHAPTER 4 BIVARIATE DISTRIBUTION EXTENSION
CHAPTER 4 BIVARIATE DISTRIBUTION EXTENSION 4. Itroductio Numerous bivariate discrete distributios have bee defied ad studied (see Mardia, 97 ad Kocherlakota ad Kocherlakota, 99) based o various methods
More informationComparison Study of Series Approximation. and Convergence between Chebyshev. and Legendre Series
Applied Mathematical Scieces, Vol. 7, 03, o. 6, 3-337 HIKARI Ltd, www.m-hikari.com http://d.doi.org/0.988/ams.03.3430 Compariso Study of Series Approimatio ad Covergece betwee Chebyshev ad Legedre Series
More informationTopic 9: Sampling Distributions of Estimators
Topic 9: Samplig Distributios of Estimators Course 003, 2016 Page 0 Samplig distributios of estimators Sice our estimators are statistics (particular fuctios of radom variables), their distributio ca be
More informationDepartment of Mathematics
Departmet of Mathematics Ma 3/103 KC Border Itroductio to Probability ad Statistics Witer 2017 Lecture 19: Estimatio II Relevat textbook passages: Larse Marx [1]: Sectios 5.2 5.7 19.1 The method of momets
More information[412] A TEST FOR HOMOGENEITY OF THE MARGINAL DISTRIBUTIONS IN A TWO-WAY CLASSIFICATION
[412] A TEST FOR HOMOGENEITY OF THE MARGINAL DISTRIBUTIONS IN A TWO-WAY CLASSIFICATION BY ALAN STUART Divisio of Research Techiques, Lodo School of Ecoomics 1. INTRODUCTION There are several circumstaces
More informationECE 8527: Introduction to Machine Learning and Pattern Recognition Midterm # 1. Vaishali Amin Fall, 2015
ECE 8527: Itroductio to Machie Learig ad Patter Recogitio Midterm # 1 Vaishali Ami Fall, 2015 tue39624@temple.edu Problem No. 1: Cosider a two-class discrete distributio problem: ω 1 :{[0,0], [2,0], [2,2],
More informationSection 14. Simple linear regression.
Sectio 14 Simple liear regressio. Let us look at the cigarette dataset from [1] (available to dowload from joural s website) ad []. The cigarette dataset cotais measuremets of tar, icotie, weight ad carbo
More informationECE 901 Lecture 12: Complexity Regularization and the Squared Loss
ECE 90 Lecture : Complexity Regularizatio ad the Squared Loss R. Nowak 5/7/009 I the previous lectures we made use of the Cheroff/Hoeffdig bouds for our aalysis of classifier errors. Hoeffdig s iequality
More informationMaximum likelihood estimation from record-breaking data for the generalized Pareto distribution
METRON - Iteratioal Joural of Statistics 004, vol. LXII,. 3, pp. 377-389 NAGI S. ABD-EL-HAKIM KHALAF S. SULTAN Maximum likelihood estimatio from record-breakig data for the geeralized Pareto distributio
More informationSimulation. Two Rule For Inverting A Distribution Function
Simulatio Two Rule For Ivertig A Distributio Fuctio Rule 1. If F(x) = u is costat o a iterval [x 1, x 2 ), the the uiform value u is mapped oto x 2 through the iversio process. Rule 2. If there is a jump
More informationMaximum Likelihood Estimation
Chapter 9 Maximum Likelihood Estimatio 9.1 The Likelihood Fuctio The maximum likelihood estimator is the most widely used estimatio method. This chapter discusses the most importat cocepts behid maximum
More informationAxis Aligned Ellipsoid
Machie Learig for Data Sciece CS 4786) Lecture 6,7 & 8: Ellipsoidal Clusterig, Gaussia Mixture Models ad Geeral Mixture Models The text i black outlies high level ideas. The text i blue provides simple
More informationRAINFALL PREDICTION BY WAVELET DECOMPOSITION
RAIFALL PREDICTIO BY WAVELET DECOMPOSITIO A. W. JAYAWARDEA Departmet of Civil Egieerig, The Uiversit of Hog Kog, Hog Kog, Chia P. C. XU Academ of Mathematics ad Sstem Scieces, Chiese Academ of Scieces,
More informationSurveying the Variance Reduction Methods
Iteratioal Research Joural of Applied ad Basic Scieces 2013 Available olie at www.irjabs.com ISSN 2251-838X / Vol, 7 (7): 427-432 Sciece Explorer Publicatios Surveyig the Variace Reductio Methods Arash
More informationStatistical Inference (Chapter 10) Statistical inference = learn about a population based on the information provided by a sample.
Statistical Iferece (Chapter 10) Statistical iferece = lear about a populatio based o the iformatio provided by a sample. Populatio: The set of all values of a radom variable X of iterest. Characterized
More informationLinear regression. Daniel Hsu (COMS 4771) (y i x T i β)2 2πσ. 2 2σ 2. 1 n. (x T i β y i ) 2. 1 ˆβ arg min. β R n d
Liear regressio Daiel Hsu (COMS 477) Maximum likelihood estimatio Oe of the simplest liear regressio models is the followig: (X, Y ),..., (X, Y ), (X, Y ) are iid radom pairs takig values i R d R, ad Y
More informationECONOMETRIC THEORY. MODULE XIII Lecture - 34 Asymptotic Theory and Stochastic Regressors
ECONOMETRIC THEORY MODULE XIII Lecture - 34 Asymptotic Theory ad Stochastic Regressors Dr. Shalabh Departmet of Mathematics ad Statistics Idia Istitute of Techology Kapur Asymptotic theory The asymptotic
More informationSDS 321: Introduction to Probability and Statistics
SDS 321: Itroductio to Probability ad Statistics Lecture 23: Cotiuous radom variables- Iequalities, CLT Puramrita Sarkar Departmet of Statistics ad Data Sciece The Uiversity of Texas at Austi www.cs.cmu.edu/
More informationMixtures of Gaussians and the EM Algorithm
Mixtures of Gaussias ad the EM Algorithm CSE 6363 Machie Learig Vassilis Athitsos Computer Sciece ad Egieerig Departmet Uiversity of Texas at Arligto 1 Gaussias A popular way to estimate probability desity
More informationLecture 7: Density Estimation: k-nearest Neighbor and Basis Approach
STAT 425: Itroductio to Noparametric Statistics Witer 28 Lecture 7: Desity Estimatio: k-nearest Neighbor ad Basis Approach Istructor: Ye-Chi Che Referece: Sectio 8.4 of All of Noparametric Statistics.
More informationStatistical Inference Based on Extremum Estimators
T. Rotheberg Fall, 2007 Statistical Iferece Based o Extremum Estimators Itroductio Suppose 0, the true value of a p-dimesioal parameter, is kow to lie i some subset S R p : Ofte we choose to estimate 0
More informationLecture 2: Monte Carlo Simulation
STAT/Q SCI 43: Itroductio to Resamplig ethods Sprig 27 Istructor: Ye-Chi Che Lecture 2: ote Carlo Simulatio 2 ote Carlo Itegratio Assume we wat to evaluate the followig itegratio: e x3 dx What ca we do?
More informationJournal of Multivariate Analysis. Superefficient estimation of the marginals by exploiting knowledge on the copula
Joural of Multivariate Aalysis 102 (2011) 1315 1319 Cotets lists available at ScieceDirect Joural of Multivariate Aalysis joural homepage: www.elsevier.com/locate/jmva Superefficiet estimatio of the margials
More informationChapter 6 Principles of Data Reduction
Chapter 6 for BST 695: Special Topics i Statistical Theory. Kui Zhag, 0 Chapter 6 Priciples of Data Reductio Sectio 6. Itroductio Goal: To summarize or reduce the data X, X,, X to get iformatio about a
More informationPattern Classification
Patter Classificatio All materials i these slides were tae from Patter Classificatio (d ed) by R. O. Duda, P. E. Hart ad D. G. Stor, Joh Wiley & Sos, 000 with the permissio of the authors ad the publisher
More information1.010 Uncertainty in Engineering Fall 2008
MIT OpeCourseWare http://ocw.mit.edu.00 Ucertaity i Egieerig Fall 2008 For iformatio about citig these materials or our Terms of Use, visit: http://ocw.mit.edu.terms. .00 - Brief Notes # 9 Poit ad Iterval
More informationGoodness-Of-Fit For The Generalized Exponential Distribution. Abstract
Goodess-Of-Fit For The Geeralized Expoetial Distributio By Amal S. Hassa stitute of Statistical Studies & Research Cairo Uiversity Abstract Recetly a ew distributio called geeralized expoetial or expoetiated
More informationThe Method of Least Squares. To understand least squares fitting of data.
The Method of Least Squares KEY WORDS Curve fittig, least square GOAL To uderstad least squares fittig of data To uderstad the least squares solutio of icosistet systems of liear equatios 1 Motivatio Curve
More informationRandom Variables, Sampling and Estimation
Chapter 1 Radom Variables, Samplig ad Estimatio 1.1 Itroductio This chapter will cover the most importat basic statistical theory you eed i order to uderstad the ecoometric material that will be comig
More informationBayesian inference for Parameter and Reliability function of Inverse Rayleigh Distribution Under Modified Squared Error Loss Function
Australia Joural of Basic ad Applied Scieces, (6) November 26, Pages: 24-248 AUSTRALIAN JOURNAL OF BASIC AND APPLIED SCIENCES ISSN:99-878 EISSN: 239-844 Joural home page: www.ajbasweb.com Bayesia iferece
More informationTHE SYSTEMATIC AND THE RANDOM. ERRORS - DUE TO ELEMENT TOLERANCES OF ELECTRICAL NETWORKS
R775 Philips Res. Repts 26,414-423, 1971' THE SYSTEMATIC AND THE RANDOM. ERRORS - DUE TO ELEMENT TOLERANCES OF ELECTRICAL NETWORKS by H. W. HANNEMAN Abstract Usig the law of propagatio of errors, approximated
More informationFirst Year Quantitative Comp Exam Spring, Part I - 203A. f X (x) = 0 otherwise
First Year Quatitative Comp Exam Sprig, 2012 Istructio: There are three parts. Aswer every questio i every part. Questio I-1 Part I - 203A A radom variable X is distributed with the margial desity: >
More informationLecture 33: Bootstrap
Lecture 33: ootstrap Motivatio To evaluate ad compare differet estimators, we eed cosistet estimators of variaces or asymptotic variaces of estimators. This is also importat for hypothesis testig ad cofidece
More informationSTA Object Data Analysis - A List of Projects. January 18, 2018
STA 6557 Jauary 8, 208 Object Data Aalysis - A List of Projects. Schoeberg Mea glaucomatous shape chages of the Optic Nerve Head regio i aimal models 2. Aalysis of VW- Kedall ati-mea shapes with a applicatio
More informationVector Quantization: a Limiting Case of EM
. Itroductio & defiitios Assume that you are give a data set X = { x j }, j { 2,,, }, of d -dimesioal vectors. The vector quatizatio (VQ) problem requires that we fid a set of prototype vectors Z = { z
More informationMathematical Modeling of Optimum 3 Step Stress Accelerated Life Testing for Generalized Pareto Distribution
America Joural of Theoretical ad Applied Statistics 05; 4(: 6-69 Published olie May 8, 05 (http://www.sciecepublishiggroup.com/j/ajtas doi: 0.648/j.ajtas.05040. ISSN: 6-8999 (Prit; ISSN: 6-9006 (Olie Mathematical
More informationCS434a/541a: Pattern Recognition Prof. Olga Veksler. Lecture 5
CS434a/54a: Patter Recogitio Prof. Olga Veksler Lecture 5 Today Itroductio to parameter estimatio Two methods for parameter estimatio Maimum Likelihood Estimatio Bayesia Estimatio Itroducto Bayesia Decisio
More informationElement sampling: Part 2
Chapter 4 Elemet samplig: Part 2 4.1 Itroductio We ow cosider uequal probability samplig desigs which is very popular i practice. I the uequal probability samplig, we ca improve the efficiecy of the resultig
More informationBootstrap Intervals of the Parameters of Lognormal Distribution Using Power Rule Model and Accelerated Life Tests
Joural of Moder Applied Statistical Methods Volume 5 Issue Article --5 Bootstrap Itervals of the Parameters of Logormal Distributio Usig Power Rule Model ad Accelerated Life Tests Mohammed Al-Ha Ebrahem
More informationBull. Korean Math. Soc. 36 (1999), No. 3, pp. 451{457 THE STRONG CONSISTENCY OF NONLINEAR REGRESSION QUANTILES ESTIMATORS Seung Hoe Choi and Hae Kyung
Bull. Korea Math. Soc. 36 (999), No. 3, pp. 45{457 THE STRONG CONSISTENCY OF NONLINEAR REGRESSION QUANTILES ESTIMATORS Abstract. This paper provides suciet coditios which esure the strog cosistecy of regressio
More informationMultivariate Analysis of Variance Using a Kotz Type Distribution
Proceedigs of the World Cogress o Egieerig 2008 Vol II WCE 2008, July 2-4, 2008, Lodo, UK Multivariate Aalysis of Variace Usig a Kotz Type Distributio Kusaya Plugpogpu, ad Dayaad N Naik Abstract Most stadard
More informationEstimation of Backward Perturbation Bounds For Linear Least Squares Problem
dvaced Sciece ad Techology Letters Vol.53 (ITS 4), pp.47-476 http://dx.doi.org/.457/astl.4.53.96 Estimatio of Bacward Perturbatio Bouds For Liear Least Squares Problem Xixiu Li School of Natural Scieces,
More informationAsymptotic Properties of MLE in Stochastic. Differential Equations with Random Effects in. the Diffusion Coefficient
Iteratioal Joural of Cotemporary Mathematical Scieces Vol. 1, 215, o. 6, 275-286 HIKARI Ltd, www.m-hikari.com http://dx.doi.org/1.12988/ijcms.215.563 Asymptotic Properties of MLE i Stochastic Differetial
More informationProperties and Hypothesis Testing
Chapter 3 Properties ad Hypothesis Testig 3.1 Types of data The regressio techiques developed i previous chapters ca be applied to three differet kids of data. 1. Cross-sectioal data. 2. Time series data.
More information11 Correlation and Regression
11 Correlatio Regressio 11.1 Multivariate Data Ofte we look at data where several variables are recorded for the same idividuals or samplig uits. For example, at a coastal weather statio, we might record
More informationResearch Article A Unified Weight Formula for Calculating the Sample Variance from Weighted Successive Differences
Discrete Dyamics i Nature ad Society Article ID 210761 4 pages http://dxdoiorg/101155/2014/210761 Research Article A Uified Weight Formula for Calculatig the Sample Variace from Weighted Successive Differeces
More informationA RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS
J. Japa Statist. Soc. Vol. 41 No. 1 2011 67 73 A RANK STATISTIC FOR NON-PARAMETRIC K-SAMPLE AND CHANGE POINT PROBLEMS Yoichi Nishiyama* We cosider k-sample ad chage poit problems for idepedet data i a
More informationJoint Probability Distributions and Random Samples. Jointly Distributed Random Variables. Chapter { }
UCLA STAT A Applied Probability & Statistics for Egieers Istructor: Ivo Diov, Asst. Prof. I Statistics ad Neurology Teachig Assistat: Neda Farziia, UCLA Statistics Uiversity of Califoria, Los Ageles, Sprig
More informationof the matrix is =-85, so it is not positive definite. Thus, the first
BOSTON COLLEGE Departmet of Ecoomics EC771: Ecoometrics Sprig 4 Prof. Baum, Ms. Uysal Solutio Key for Problem Set 1 1. Are the followig quadratic forms positive for all values of x? (a) y = x 1 8x 1 x
More informationRank tests and regression rank scores tests in measurement error models
Rak tests ad regressio rak scores tests i measuremet error models J. Jurečková ad A.K.Md.E. Saleh Charles Uiversity i Prague ad Carleto Uiversity i Ottawa Abstract The rak ad regressio rak score tests
More informationRegression with an Evaporating Logarithmic Trend
Regressio with a Evaporatig Logarithmic Tred Peter C. B. Phillips Cowles Foudatio, Yale Uiversity, Uiversity of Aucklad & Uiversity of York ad Yixiao Su Departmet of Ecoomics Yale Uiversity October 5,
More informationApproximating the ruin probability of finite-time surplus process with Adaptive Moving Total Exponential Least Square
WSEAS TRANSACTONS o BUSNESS ad ECONOMCS S. Khotama, S. Boothiem, W. Klogdee Approimatig the rui probability of fiite-time surplus process with Adaptive Movig Total Epoetial Least Square S. KHOTAMA, S.
More informationRegression and generalization
Regressio ad geeralizatio CE-717: Machie Learig Sharif Uiversity of Techology M. Soleymai Fall 2016 Curve fittig: probabilistic perspective Describig ucertaity over value of target variable as a probability
More informationCSE 527, Additional notes on MLE & EM
CSE 57 Lecture Notes: MLE & EM CSE 57, Additioal otes o MLE & EM Based o earlier otes by C. Grat & M. Narasimha Itroductio Last lecture we bega a examiatio of model based clusterig. This lecture will be
More informationProbability and MLE.
10-701 Probability ad MLE http://www.cs.cmu.edu/~pradeepr/701 (brief) itro to probability Basic otatios Radom variable - referrig to a elemet / evet whose status is ukow: A = it will rai tomorrow Domai
More informationDimension-free PAC-Bayesian bounds for the estimation of the mean of a random vector
Dimesio-free PAC-Bayesia bouds for the estimatio of the mea of a radom vector Olivier Catoi CREST CNRS UMR 9194 Uiversité Paris Saclay olivier.catoi@esae.fr Ilaria Giulii Laboratoire de Probabilités et
More informationConfidence Interval for Standard Deviation of Normal Distribution with Known Coefficients of Variation
Cofidece Iterval for tadard Deviatio of Normal Distributio with Kow Coefficiets of Variatio uparat Niwitpog Departmet of Applied tatistics, Faculty of Applied ciece Kig Mogkut s Uiversity of Techology
More informationAl- Mustansiriyah J. Sci. Vol. 24, No 5, 2013
Al- Mustasiriyah J. Sci. Vol. 24, No 5, 23 Usig Etropy Loss Fuctio to Estimate the Scale Parameter for Laplace Distributio Huda A. Rasheed, Akbal J. Sulta ad Nadia J. Fazah Departmet of Mathematics, college
More information