DOCSLIB.ORG

Parameter Estimation for Tweedie Distribution with MCMC Approach Topic Presentation

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Parameter Estimation for Tweedie Distribution with MCMC Approach Topic Presentation Parameter Estimation for Tweedie Distribution with MCMC Approach Topic Presentation Tairan Ye 4/12/2018 University of Connecticut Summary 1. Markov Chain Monte Carlo 2. Tweedie Distribution 3. Parameter Estimation for Tweedie Model with MCMC Approach 1 Markov Chain Monte Carlo Introduction • Fact: Your desired parameters are not in the analytical form or extremely hard to solve out. • Solution: The sampling algorithms, like Markov Chain Monte Carlo • Reason: By constructing a Markov chain, we can sample the desired distribution by observing the chain after a number of iterations. The sample distribution will match more closely to the actual desired distribution as the iterations increase. 2 Example: Monte Carlo Integration General problem: Evaluating E[h(X )] = R h(x)p(x)dx can be difficult. However, if we can draw samples: X (1), X (2), ..., X (N) ∼ p(x) (1) Then, we can estimate: 1 N E[h(X )] ≈ ∑ h(X (t)) (2) N t=1 This is Monte Carlo (MC) Integration. 3 Markov Chain A Markov chain is generated by sampling: X (t+1) ∼ p(xjx(t)), t = 1, 2, ... (3) Here p is the transition kernel. So X (t+1) depends only on X (t), rather than X (0), X (1), X (2), ..., X (t−1) We can summarize as following: p(X (t+1)jX (t), X (t−1), ..., X (0)) = p(X (t+1)jX (t)) (4) 4 An example of Markov Chain A first order auto-regressive process with lag-1 auto-correlation 0.5. X (t+1)jX (t) ∼ N(0.5x(t), 1) (5) A simulation is conducted as following, which contains two different starting points: Figure 1: Simulation with two starting points The chains seemed to have forgotten their starting positions. 5 Sampling Algorithm Overview The most fundamental MCMC sampling algorithm are: • Gibbs Sampling 1. Require for the analytic form of the full conditional distribution. 2. High-efficiency { the sequence will converge with little iterations. • Metropolis-Hasting Algorithm 1. More flexible { no need for the deduction of full conditional distribution. 2. Some challenges for setting up the proposal distribution { experience needed to control the accept rate between 20% to 30%. 6 Full Conditional Distribution and Gibbs Sampling • Full conditional distributions: 2 2 The distributions p(qjs , y1, ..., yn) and p(s jq, y1, ..., yn) are called the full conditional distributions of s2 and q respectively, as they are each a conditional distribution of a parameter given everything else. • Gibbs Sampler: Given a current state of the parameters f(s) = fq(s), s2(s)g, we generate a new state as follows: (s+1) 2(s) 1. sample q ∼ p(qjs , y1, ..., yn); 2(s+1) 2 (s+1) 2. sample s ∼ p(s jq , y1, ..., yn); 3. update f(s+1) = fq(s+1), s2(s+1)g 7 Generalized Gibbs Sampler Suppose you have a vector of parameters f = ff1, ..., fpg, and your information about f is measured with p(f) = p(f1, ..., fp). For example, in the normal model f = fq, s2g, and the probability measure of interest 2 (0) (0) (0) is p(q, s jy1, ..., yn). Given a starting point f = ff1 , ..., fp g, the Gibbs sampler generates f(s) from f(s−1) as follows: (s) (s−1) (s−1) (s−1) • sample f1 ∼ p(f1jf2 , f3 , ..., fp ) (s) (s) (s−1) (s−1) • sample f2 ∼ p(f2jf1 , f3 , ..., fp ) ...... (s) (s) (s) (s) • sample fp ∼ p(fpjf1 , f2 , ..., fp−1) Markov Property: In this sequence, f(s) depends on f(0), ..., f(s−1) only through f(s−1), i.e. f(s) is conditionally independent of f(0), ..., f(s−2) given f(s−1). 8 Metropolis Algorithm The Metropolis algorithm proceeds by sampling a proposal value q∗ nearby the current value q(s) using a symmetric proposal distribution J(q∗jq(s)). Usually J(q∗jq(s)) is very simple, with samples from J(q∗jq(s)) being near q(s) with high probability. Examples include: • J(q∗jq(s)) ∼ U(q(s) − d, q(s) + d) • J(q∗jq(s)) ∼ N(q(s), d2) 9 Metropolis Algorithm The algorithm is as following: 1. Sample q∗ ∼ J(q∗jq(s)) 2. Compute the acceptance ratio: p(q∗jy) p(yjq∗)p(q∗) r = = (6) p(q(s)jy) p(yjq(s))p(q(s)) 3. Update ( ∗ q , p = min(r, 1) q(s+1) = (7) q(s), p = 1 − min(r, 1) • Step 3 can be accomplished by sampling u ∼ U(0, 1) and setting q(s+1) = q∗ if u < r and setting q(s+1) = q(s) otherwise. • In many cases, computing the ratio r directly can be numerically unstable, a problem that often can be remedied by computing the logarithm of r: 10 Metropolis Hastings Algorithm Metropolis Hastings Algorithm is the generalized and multi-parameters version of Gibbs sampling and Metropolis algorithm. Let's consider a simple example where our target probability distribution is p0(u, v), a bivariate distribution for two random variables U and V . 1. Update U: ∗ ∗ (s) (s) • Sample u ∼ Ju(u ju , v ) • Compute the acceptance ratio: p (u∗, v (s)) J (u(s)ju∗, v (s)) r = 0 ∗ u (8) (s) (s) ∗ (s) (s) p0(u , v ) Ju(u ju , v • Update ( u∗, p = min(r, 1) u(s+1) = (9) u(s), p = 1 − min(r, 1) • Sampling u ∼ U(0, 1) and setting u(s+1) = u∗ if u < r and setting u(s+1) = u(s) otherwise. 11 Metropolis Hastings Algorithm After updating U, we continue to update V: 1. Update V : ∗ ∗ (s+1) (s) • Sample v ∼ Jv (v ju , v ) • Compute the acceptance ratio: p (u(s+1), v ∗) J (v (s)ju(s+1), v ∗) r = 0 ∗ v (10) (s+1) (s) ∗ (s+1) (s) p0(u , v ) Jv (v ju , v ) • Update ( v ∗, p = min(r, 1) v (s+1) = (11) v (s), p = 1 − min(r, 1) • Sampling u ∼ U(0, 1) and setting v (s+1) = v ∗ if u < r and setting v (s+1) = v (s) otherwise. 12 Proposal Distribution The proposal distribution is the most "challenging" part for conducting the Metropolis algorithm. • Without restriction: ∗ 2 Normal distribution: q ∼ N(qt−1, sp ) • With restriction: ∗ Random walk: f (q ) = f (qt−1) + e and e ∼ N(0, 1) Example: q > 0, then apply log function. ∗ 2 q can be generated from lognormal(log(qt−1), sp ) 13 Combination In complex models, it is often the case that conditional distributions are available for some parameters but not for others. In these situations we can combine Gibbs and Metropolis-type proposal distributions to generate a Markov chain to approximate the joint distribution of all of the parameters. 14 More to talk about 1. Stationary: As t ! ¥, the Markov chain converges to its stationary distribution. Technique: trace plot, effective sample size (ESS) Package: 'coda' in R 2. Burn-in: Typically, we burn-in the initial simulation values since they are most likely nonstationary, and study the remaining part. 3. Gap: In order to avoid the dependency, we need to pick up the simulation values by gaps. 4. Chains: Also, we will build more than 1 chain to make sure the sampler is representative for the distribution. By the way, the calculation of deviance information criterion (DIC) requires at least 2 chains. 15 Example Figure 2: The trace plot of the raw MCMC 16 Example Figure 3: The trace plot of the updated sequence 17 Software • WinBUGS (Bayesian inference Using Gibbs Sampling) • JAGS (Just Another Gibbs Sampler) • Stan { with linkage of most of data science softwares via multiply platforms • R packages: rjags, coda, mcmc, BRugs, etc. 18 Tweedie Distribution Tweedie Distribution Overview Tweedie distribution can be defined as Tw(m, f, p). • m is the expected mean for the Tweedie distribution. • f is called the dispersion parameter. • p, or power, controls what distribution it belongs to. Property: • A special case of exponential dispersion models. • Variance = fmp. • Most interested in p 2 (1, 2), which is defined as compound Poisson-gamma distribution. 19 Compound Poisson-gamma distribution Due to its special property { mass at zero and continue otherwise, the Tweedie model with p 2 (1, 2), or the compound Poisson-gamma distribution, is popular in insurance industry. Denote the total auto insurance claim in dollar is Y , the frequency of claim is N, and the severity of each claim is Z. Then, we have: N Y = ∑ Zn (12) n=1 N ∼ Poisson(l) (13) Z ∼ gamma(a, b) (14) N = Z (15) j 20 Compound Poisson-gamma distribution We say Y ∼ Tw(m, f, p). Also: lab = m (16) la(a + 1)b2 = fmp (17) (18) We can solve: m2−p l = (19) f(2 − p) 2 − p a = (20) p − 1 b = f(p − 1)mp−1 (21) 21 Compound Poisson-gamma distribution The probability function for the Tweedie distribution with a power p 2 (1, 2) is: 1. For y = 0: P[Y = 0] = exp f−lg m2−p (22) = exp − f(2 − p) 2. For y > 0: ¥ n − y 1 l f (y) = e b e−l y na−1 Y ∑ na n=1 G(na)b n! j n 1 y 2−p o ( (2−p) ) ¥ ( ) p−1 y m 1 2−p f(p−1) = exp − − f(p − 1)m(p−1) f(2 − p) y ∑ j (2−p)j j=1 f G( p−1 )j! (23) 22 Parameter Estimation for Tweedie Model with MCMC Approach Model In order to study the key factors that affects the auto insurance claims, we can set up: log(ms ) = Xsg + erandom (24) 2 erandom ∼ N(0, s ) (25) where s indicates for different location; Xs contains the information about policyholders, including personal information as well as vehicle information; g is the factors, which we are interested in; e covers all other information we don't discuss here.
Load more

Recommended publications
  • Adversarial Variational Bayes Methods for Tweedie Compound Poisson Mixed Models
    ADVERSARIAL VARIATIONAL BAYES METHODS FOR TWEEDIE COMPOUND POISSON MIXED MODELS Yaodong Yang∗, Rui Luo, Yuanyuan Liu American International Group Inc. ABSTRACT 1.5 w The Tweedie Compound Poisson-Gamma model is routinely 1.0 λ =1.9, α =49, β =0.03 random i effect used for modeling non-negative continuous data with a discrete x bi Σb Density probability mass at zero. Mixed models with random effects ac- i 0.5 U count for the covariance structure related to the grouping hier- λi αi βi archy in the data. An important application of Tweedie mixed 0.0 i yi models is pricing the insurance policies, e.g. car insurance. 0 5 10 n Tweedie Value M However, the intractable likelihood function, the unknown vari- ance function, and the hierarchical structure of mixed effects (a) Tweedie Density Function (b) Tweedie Mixed-Effects model. have presented considerable challenges for drawing inferences Fig. 1: Graphical model of Tweedie mixed-effect model. on Tweedie. In this study, we tackle the Bayesian Tweedie i:i:d mixed-effects models via variational inference approaches. In Poisson(λ);Gi ∼ Gamma(α; β): Tweedie is heavily used particular, we empower the posterior approximation by im- for modeling non-negative heavy-tailed continuous data with plicit models trained in an adversarial setting. To reduce the a discrete probability mass at zero (see Fig. 1a). As a result, variance of gradients, we reparameterize random effects, and Tweedie gains its importance from multiple domains [4, 5], in- integrate out one local latent variable of Tweedie. We also cluding actuarial science (aggregated loss/premium modeling), employ a flexible hyper prior to ensure the richness of the ap- ecology (species biomass modeling), meteorology (precipi- proximation.
    [Show full text]
  • Multivariate Tweedie Lifetimes: the Impact of Dependence
    Multivariate Tweedie Lifetimes: The Impact of Dependence Daniel H. Alai1 Zinoviy Landsman2 Michael Sherris3 CEPAR, Risk and Actuarial Studies, Australian School of Business UNSW, Sydney NSW 2052, Australia Departmant of Statistics, University of Haifa Mount Carmel, Haifa 31905, Israel DRAFT ONLY DO NOT CIRCULATE WITHOUT AUTHORS' PERMISSION Abstract Systematic improvements in mortality increases dependence in the survival distributions of insured lives. This is not accounted for in standard life tables and actuarial models used for annuity pricing and reserving. Furthermore, systematic longevity risk undermines the law of large numbers; a law that is relied on in the risk management of life insurance and annuity portfolios. This paper applies a multivari- ate Tweedie distribution to incorporate dependence, which it induces through a common shock component. Model parameter estimation is developed based on the method of moments and generalized to allow for truncated observations. Keywords: systematic longevity risk, dependence, multivariate Tweedie, lifetime distribution [email protected] [email protected] [email protected] 1 Introduction The paper generalizes an approach explored in a previous paper by the au- thors to model the dependence of lifetimes using the gamma distribution with a common stochastic component. The property of gamma random variables generates a multivariate gamma distribution using the so-called multivari- ate reduction method; see Chereiyan (1941) and Ramabhadran (1951). This constructs a dependency structure that is natural for modelling lifetimes of individuals within a pool. The method uses the fact that a sum of gamma random variables with the same rate parameter follows a gamma distribution with that rate parameter.
    [Show full text]
  • Robust Monitoring of Time Series with Application to Fraud Detection
    Robust Monitoring of Time Series with Application to Fraud Detection Peter Rousseeuw, Domenico Perrotta, Marco Riani and Mia Hubert May 22, 2018 Abstract Time series often contain outliers and level shifts or structural changes. These unexpected events are of the utmost importance in fraud detection, as they may pin- point suspicious transactions. The presence of such unusual events can easily mislead conventional time series analysis and yield erroneous conclusions. A unified frame- work is provided for detecting outliers and level shifts in short time series that may have a seasonal pattern. The approach combines ideas from the FastLTS algorithm for robust regression with alternating least squares. The double wedge plot is pro- posed, a graphical display which indicates outliers and potential level shifts. The methodology was developed to detect potential fraud cases in time series of imports into the European Union, and is illustrated on two such series. Keywords: alternating least squares, double wedge plot, level shift, outliers. arXiv:1708.08268v4 [stat.CO] 20 May 2018 Peter Rousseeuw, Department of Mathematics, University of Leuven, Belgium. Domenico Perrotta, Joint Research Centre, Ispra, Italy. Marco Riani, Department of Economics, University of Parma, Italy. Mia Hubert, Department of Mathematics, University of Leuven, Belgium. Peter Rousseeuw and Mia Hubert gratefully acknowledge the support by project C16/15/068 of Internal Funds KU Leuven. The work of Domenico Perrotta was supported by the Project “Automated Monitoring Tool on External Trade Step 5” of the Joint Research Centre and the European Anti-Fraud Office of the European Commission, under the Hercule-III EU programme. In this paper, references to specific countries and products are made only for purposes of illustration and do not necessarily refer to cases investigated or under investigation by anti-fraud authorities.
    [Show full text]
  • Generalized Linear Models Lecture 11
    Generalized Linear Models Lecture 11. Tweedie models. Compound Poisson models GLM (MTMS.01.011) Lecture 11 1 / 22 Tweedie distributions Tw(µ, p, ϕ), Twp(µ, ϕ) Tweedie distributions form a subclass of exponential dispersion family, the class includes continuous distributions like normal, gamma and inverse gaussian, discrete distributions like Poisson, and also Poisson-gamma mixtures (that have positive probability at zero and are continuous elsewhere) The variance function for Tweedie distributions has the following form ν(µ) = µp GLM with Tweedie response Canonical parameter θi and canonical function b(θi ) are ( 1−p ( 2−p µi µ(θi ) 1−p , p 6= 1 2−p , p 6= 2 θi = θ(µi ) = b(θi ) = log µi , p = 1 log µ(θi ), p = 2 Mean: µi p Variance: ϕµi – power-variance distributions The Tweedie distributions were named by Bent Jorgensen after Maurice Tweedie, a statistician and medical physicist at the University of Liverpool, UK, who presented the first thorough study of these distributions in 1984. GLM (MTMS.01.011) Lecture 11 2 / 22 Tweedie family Twp(µ, ϕ) Tweedie distributions exist for all p > 0, no analytic form exists for 0 < p < 1 If 1 < p < 2, the distribution are continuous for Y > 0, with a positive mass at Y = 0. If p > 2, the distributions are continuous, Y > 0 Known distributions: p = 0 – normal distribution p = 1 and ϕ = 1 – Poisson distribution 1 < p < 2 – compound Poisson-gamma distribution p = 2 – gamma distribution p = 3 – inverse gaussian distribution 2 < p < 3,p > 3 – positive stable distributions There are Tweedie models that allow for zero-issues as well GLM (MTMS.01.011) Lecture 11 3 / 22 Applications of Twp distribution 1 1 < p < 2 Fish count estimation (silky shark, tuna fish): pˆ = 1.12, Shono (2008, 2010) Root length density of apple trees: pˆ = 1.4, Silva (1999) 2 p ≥ 2 Tweedie distribution with p ≥ 2 is a continuous non-negative distribution.
    [Show full text]
  • Modelling Rainfall Using Generalized Linear Models When the Response Variable Has a Tweedie Distribution
    MODELLING RAINFALL USING GENERALIZED LINEAR MODELS WHEN THE RESPONSE VARIABLE HAS A TWEEDIE DISTRIBUTION By Kanithi.K.Isaiah 756/ 8084/2006 UNIVERSITY OF NAIROBI COLLEGE OF BIOLOGICAL AND PHYSICAL SCIENCES SCHOOL OF MATHEMATICS A project submitted in partial fulfillment of the requirements for the Degree of Master of Science in Social Statistics at th e University of Nairobi August 2009 University of N A IRO B I Library Declaration This project is iny original work and has not been presented for a degree award in any other University. "Pi Signatu D ato^.-.A^Sr..^ Kanithi.K .Isaiah This Project has been submitted for examination with my approval as a University supervisor. Acknowledgements The following people have assisted me in the succesful completion of this project.I thank yon all and will eternally be grateful. • My supervisor Dr.T.O.Achia,University of Nairobi,who guided and in­ spired me throughout the course of this project.He also spent long hours in managing the huge dataset that we eventually worked with. • Dr.Christopher Oludhe of The Department of Meteorology,who pro­ vided the data and assisted me with reference materials.He was also readily available for any explanations or information that I required. • Mr. A.Njogu of Dagoretti Meteorological station who took time to explain to me the various climatological issues and introduced me to Instat-f. • Last but not least are Proffessor Biamah and Dr.Gichuki of The Depart­ ment of Environmental Engineering,School of Engineering,University of Nairobi for their support. ii t Dedication This project is dedicated to my family for the continued sup­ port that they have given me over the years as I pursued my education.GOD bless YOU ALL.
    [Show full text]
  • Evaluation of Tweedie Exponential Dispersion Model Densities by Fourier Inversion
    View metadata, citation and similar papers at core.ac.uk brought to you by CORE provided by University of Southern Queensland ePrints Evaluation of Tweedie exponential dispersion model densities by Fourier inversion Peter K. Dunn e-mail: [email protected] Australian Centre for Sustainable Catchments and Department of Mathematics and Computing University of Southern Queensland Toowoomba Queensland 4350 Australia Gordon K. Smyth Bioinformatics Division Walter and Eliza Hall Institute of Medical Research Melbourne, Vic 3050, Australia May 9, 2007 Abstract The Tweedie family of distributions is a family of exponential dispersion models with power variance functions V (µ) = µp for p (0, 1). These distri- 6∈ butions do not generally have density functions that can be written in closed form. However, they have simple moment generating functions, so the densities can be evaluated numerically by Fourier inversion of the characteristic func- tions. This paper develops numerical methods to make this inversion fast and accurate. Acceleration techniques are used to handle oscillating integrands. A range of analytic results are used to ensure convergent computations and to reduce the complexity of the parameter space. The Fourier inversion method is compared to a series evaluation method and the two methods are found to be complementary in that they perform well in different regions of the parameter space. Keywords: compound Poisson distribution; generalized linear models; nu- merical integration; numerical acceleration; power variance function 1 Introduction It is well known that the density function of a statistical distribution can be represented as an integral in terms of the characteristic function for that distribution (Abramowitz and Stegun, 1965, 26.1.10).
    [Show full text]
  • How Much Can a Retailer Sell? Sales Forecasting on Tmall
    How Much Can A Retailer Sell? Sales Forecasting on Tmall Chaochao Chen?, Ziqi Liu?, Jun Zhou, Xiaolong Li, Yuan Qi, Yujing Jiao, and Xingyu Zhong Ant Financial Services Group Hangzhou, China, 310099 fchaochao.ccc, ziqiliu, jun.zhoujun, xl.li, yuan.qi, yujing.jyj, [email protected] Abstract. Time-series forecasting is an important task in both aca- demic and industry, which can be applied to solve many real forecasting problems like stock, water-supply, and sales predictions. In this paper, we study the case of retailers' sales forecasting on Tmall|the world's leading online B2C platform. By analyzing the data, we have two main observations, i.e., sales seasonality after we group different groups of re- tails and a Tweedie distribution after we transform the sales (target to forecast). Based on our observations, we design two mechanisms for sales forecasting, i.e., seasonality extraction and distribution transformation. First, we adopt Fourier decomposition to automatically extract the sea- sonalities for different categories of retailers, which can further be used as additional features for any established regression algorithms. Second, we propose to optimize the Tweedie loss of sales after logarithmic trans- formations. We apply these two mechanisms to classic regression models, i.e., neural network and Gradient Boosting Decision Tree, and the ex- perimental results on Tmall dataset show that both mechanisms can significantly improve the forecasting results. Keywords: sales forecasting · Tweedie distribution · distribution trans- form · seasonality extraction. 1 Introduction Time-series forecasting is an important task in both academic [4] and industry arXiv:2002.11940v1 [cs.LG] 27 Feb 2020 [16], which can be applied to solve many real forecasting problems including stock, water-supply, and sales predictions.
    [Show full text]
  • Generalized Estimating Equations When the Response Variable Has a Tweedie Distribution: an Application for Multi-Site Rainfall Modelling
    CORE Metadata, citation and similar papers at core.ac.uk Provided by University of Southern Queensland ePrints Generalized estimating equations when the response variable has a Tweedie distribution: An application for multi-site rainfall modelling Taryn Swan Department of Mathematics and Computing The University of Southern Queensland, Toowoomba, QLD July 7, 2006 2 Abstract Improvements in computer power, data gathering techniques and an in- creased understanding of atmospheric dynamics has lead to the availability of extensive sets of rainfall data. With these improvements has come a wide- spread influx of researchers attempting to adequately model rainfall. This has proved quite difficult due to the extreme variability and uncertainty of rainfall. However, as rainfall has a major influence on virtually all human activity, modelling this seemingly unpredictable variable is extremely impor- tant. The considerable variation of rainfall is one of the main difficulties when modelling rainfall. A second difficulty is that rainfall is a continuous variable with an exact zero. Finally rainfall is a highly skewed variable and there is a general consensus that it has a non-independent structure. This dissertation will focus on modelling rainfall processes by using generalized es- timating equations, when the assumed distribution of the response variable is from the Tweedie family of distributions. The Tweedie distribution allows the different components of rainfall to be modelled simultaneously. When the Tweedie distributions are incorporated into generalized estimating equa- tion estimation techniques, not only can the non-independent structure of data be incorporate but also multiple rainfall sites can be modelled concur- rently. This dissertation will attempt to demonstrate the potential benefits of using generalized estimating equations for modelling and interpreting his- torical rainfall records by the examination of of monthly rainfall models at Emerald, Toowoomba and Gatton.
    [Show full text]
  • Generalized Linear Models Beyond the Exponential Family with Loss Reserve Applications
    Generalized Linear Models beyond the Exponential Family with Loss Reserve Applications Gary G Venter Guy Carpenter, LLC [email protected] Abstract The formulation of generalized linear models in Loss Models is a bit more general than is often seen, in that the residuals are not restricted to following a member of the exponential family. Some of the distributions this allows have potentially useful applications. The cost is that there is no longer a single form for the likeli- hood function, so each has to be fit directly. Here the use of loss distributions (frequency, severity and aggregate) in generalized linear models is addressed, along with a few other possibilities. 1 Generalized Linear Models beyond the Exponential Family with Loss Reserve Applications 1 Introduction The paradigm of a linear model is multiple regression, where the dependent variables are linear combinations of independent variables plus a residual term, which is from a single mean-zero normal distribution. Generalized linear mod- els, denoted here as GLZ1, allow nonlinear transforms of the regression mean as well as other forms for the distribution of residuals. Since many actuarial applications of GLZ are to cross-classified data, such as in a loss development triangle or classification rating plan, a two dimensional array of independent observations will be assumed, with a typical cell’s data denoted as qw,d. That is easy to generalize to more dimensions or to a single one. Klugman, Panjer and Willmot (2004) provide a fairly general definition of GLZs. To start with, let zw,d be the row vector of covariate observations for the w, d cell and β the column vector of coefficients.
    [Show full text]
  • Flexible Tweedie Regression Models for Continuous Data
    Journal of Statistical Computation and Simulation ISSN: 0094-9655 (Print) 1563-5163 (Online) Journal homepage: http://www.tandfonline.com/loi/gscs20 Flexible Tweedie regression models for continuous data Wagner Hugo Bonat & Célestin C. Kokonendji To cite this article: Wagner Hugo Bonat & Célestin C. Kokonendji (2017) Flexible Tweedie regression models for continuous data, Journal of Statistical Computation and Simulation, 87:11, 2138-2152, DOI: 10.1080/00949655.2017.1318876 To link to this article: http://dx.doi.org/10.1080/00949655.2017.1318876 View supplementary material Published online: 23 Apr 2017. Submit your article to this journal Article views: 92 View related articles View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=gscs20 Download by: [CAPES] Date: 26 May 2017, At: 09:21 JOURNAL OF STATISTICAL COMPUTATION AND SIMULATION, 2017 VOL. 87, NO. 11, 2138–2152 https://doi.org/10.1080/00949655.2017.1318876 Flexible Tweedie regression models for continuous data Wagner Hugo Bonat a and Célestin C. Kokonendjib aDepartment of Statistics, Paraná Federal University, Curitiba, Brazil; bLaboratoire de Mathématiques de Besançon, Université Bourgogne Franche-Comté, Besançon, France ABSTRACT ARTICLE HISTORY Tweedie regression models (TRMs) provide a flexible family of distributions Received 22 September 2016 to deal with non-negative right-skewed data and can handle continuous Accepted 10 April 2017 data with probability mass at zero. Estimation and inference of TRMs based KEYWORDS on the maximum likelihood (ML) method are challenged by the presence Semi-continuous data; of an infinity sum in the probability function and non-trivial restrictions on quasi-likelihood; the power parameter space.
    [Show full text]
  • Inferencing Tweedie Compound Poisson Mixed Models With
    Inferring Tweedie Compound Poisson Mixed Models with Adversarial Variational Methods Yaodong Yang, Rui Luo, Reza Khorshidi, Yuanyuan Liu American International Group Inc.∗ Abstract The Tweedie Compound Poisson-Gamma model is routinely used for modelling non-negative continuous data with a discrete probability mass at zero. Mixed models with random effects account for the covariance structure related to the grouping hierarchy in the data. An important application of Tweedie mixed models is estimating the aggregated loss for insurance policies. However, the intractable likelihood function, the unknown variance function, and the hierarchical structure of mixed ef- fects have presented considerable challenges for drawing inferences on Tweedie. In this study, we tackle the Bayesian Tweedie mixed-effects models via variational approaches. In particular, we empower the poste- rior approximation by implicit models trained in an adversarial setting. To reduce the variance of gradients, we reparameterize random effects, and integrate out one local latent variable of Tweedie. We also employ a flexible hyper prior to ensure the richness of the approximation. Our method is evaluated on both simulated and real-world data. Results show that the proposed method has smaller estimation bias on the random effects compared to traditional inference methods including MCMC; it also achieves a state-of-the-art predictive performance, meanwhile offering a richer estimation of the variance function. 1 Introduction Tweedie models [1, 2] are special members in the exponential dispersion family; they specify a power-law relationship between the variance and the mean: Var(Y ) = E(Y )P . For arbitrary positive P, the index parameter of the variance function, outside the interval of (0; 1), Tweedie corresponds to a particular stable distribution, e.g., Gaussian (P = 0), Poisson (P = 1), Gamma (P = 2), Inverse Gaussian (P = 3).
    [Show full text]
  • Fitting Tweedie's Compound Poisson-Gamma Mixture Model By
    Fitting Tweedie’s Compound Poisson-Gamma Mixture Model by Using PROC HPGENSELECT Overview Semicontinuous random variables are characterized by a continuous distribution that has point masses at one or more locations. One way to model semicontinuous data is to fit a generalized linear model by using a Tweedie distribution for the response variable. Tweedie distributions have been used in such diverse fields as actuarial science, economics, telecommunications, ecology, medicine, and meteorology. This example illustrates how to fit a Tweedie model to aggregate insurance claims payments data by using the HPGENSELECT procedure which is available in SAS/STAT 12.3, recently released with SAS 9.4. The SAS source code for this example is available as an attachment in a text file. In Adobe Acrobat, right-click the icon and select Save Embedded File to Disk. You can also double-click the icon to open the file immediately. Analysis Exponential dispersion models are the response distributions for generalized linear models. Any exponential dispersion model can be characterized by its variance function V ./, which describes the mean-variance relationship of the distribution when the dispersion is held constant. If Y follows an exponential dispersion model distribution that has mean , variance function V ./, and dispersion , then the variance of Y can be written as V .Y / V ./ D Tweedie distributions are a special case of the exponential dispersion family for which V ./ p and D V .Y / p (Dunn and Smyth 2005). The distribution is defined for all values of p except values of p in D the open interval .0; 1/. Many important known distributions are a special case of Tweedie distributions including normal (p 0), Poisson (p 1), gamma (p 2), and inverse Gaussian (p 3).
    [Show full text]