A comparison of the Lee-Carter model and AR-ARCH model for forecasting mortality rates

With the decline in the mortality level of populations, national social security systems and insurance companies of most developed countries are reconsidering their mortality tables taking into account the longevity risk. The Lee and Carter model is the first discrete-time stochastic model to consider the increased life expectancy trends in mortality rates and is still broadly used today. In this paper, we propose an alternative to the Lee-Carter model: an AR(1)-ARCH(1) model. More specifically, we compare the performance of these two models with respect to forecasting age-specific mortality in Italy. We fit the two models, with Gaussian and t-student innovations, for the matrix of Italian death rates from 1960 to 2003. We compare the forecast ability of the two approaches in out-of-sample analysis for the period 2004-2006 and find that the AR(1)-ARCH(1) model with t-student innovations provides the best fit among the models studied in this paper.     

On the continuity of the maximum in parametric linear programming

For the widest class of parametric linear programs with continuous dependence of coefficients on parameters, the following theorem is proven: for any parameter vectort ₀ in the domain of definition of the maximum, if the set of optimal solutions is bounded, then the maximum is upper semicontinuous att ₀. If the same proviso is met also in the dual program, then the maximum must be continuous att ₀.     

Parametric scaling of mathematical models

Parametric scaling, the process of extrapolation of a modelling result to new parametric conditions, is often required in model optimization, and can be important if the effects of parametric uncertainty on model predictions are to be quantified. Knowledge of the functional relationship between the model solution (y) and the system parameters (α) may also provide insight into the physical system underlying the model. This paper examines strategies for parametric scaling, assuming that only the nominal model solution y(α) and the associated parametric sensitivity coefficients (?y/?α, ?²y/?α², etc.) are known. The truncated Taylor series is shown to be a poor choice for parametric scaling, when y has known bounds. Alternate formulae are proposed which ‘build-in’ the constraints on y, thus expanding the parametric region in which the extrapolation may be valid. In the case where y has a temporal as well as a parametric dependence, the extrapolation may be further improved by removing from the Taylor series coefficients the ‘secular’ components, which refer to changes in the time scale of y(t), not to changes in y as a function of α.     

On periodically isotonic climate change

In a view of the past two millennia, is the recent 160-year global warmth an anomaly or a continuation of past climate pattern? In this work, we present two non-parametric χ² tests based on the U-test for the isotonic change-point problem proposed in an unpublished work by G. Shen and H. Xu (henceforth abbreviated as SX (unpublished)) and apply them to the global warming data. Our tests do not require any structure assumption on the underlying the deterministic mean process {μ_t} under the periodically isotonic alternative, but accommodate a large class of the models for the random error process. Being more attractive, the two tests obviate estimation of {μ_t} and possess explicit asymptotic distributions for short-range dependence data, and have asymptotic power 1 even under the local alternative.     

Characterization of dependence of multidimensional Lévy processes using Lévy copulas

This paper suggests Lévy copulas in order to characterize the dependence among components of multidimensional Lévy processes. This concept parallels the notion of a copula on the level of Lévy measures. As for random vectors, a version of Sklar's theorem states that the law of a general multivariate Lévy process is obtained by combining arbitrary univariate Lévy processes with an arbitrary Lévy copula. We construct parametric families of Lévy copulas and prove a limit theorem, which indicates how to obtain the Lévy copula of a multivariate Lévy process X from the ordinary copula of the random vector X_t for small t.     

On the perturbation of the observability equation in linear control systems

The aim of this paper is to investigate stability and sensitivity of the observability variable in linear control systems, (LCS) for short. We first present two results of Hölder continuity in the abstract framework of the ordinary differential equation initial-value problem x′(t) = f(t,x(t)),x(t ₀) = x ₀. Afterwards, we apply our results to automatic systems, providing henceforth the sharpest bounds for the parametric input-output relation in LCS.     

Heavy-traffic extreme value limits for Erlang delay models

We consider the maximum queue length and the maximum number of idle servers in the classical Erlang delay model and the generalization allowing customer abandonment—the M/M/n+M queue. We use strong approximations to show, under regularity conditions, that properly scaled versions of the maximum queue length and maximum number of idle servers over subintervals [0,t] in the delay models converge jointly to independent random variables with the Gumbel extreme value distribution in the quality-and-efficiency-driven (QED) and ED many-server heavy-traffic limiting regimes as n and t increase to infinity together appropriately; we require that t _n →∞ and t _n =o(n ^1/2?ε ) as n→∞ for some ε>0.     

Une méthode semi-paramétrique pour tester un modèle de régression

The purpose is to test the hypothesis H₀ that a regression model is parametric and belongs to a given family versus the alternative H₁ approaches the hypothesis from a specific direction at the rate n^?1/2. For that, we consider an empirical process such that under H₀ this process depends of a parameter θ₀. First, we start by estimating the parameter and we prove that the empirical process converges in distribution to a certain Gaussian process when the parameter is replaced by its estimator $\text{θ}{}_{\mathrm{n}}\text{.}$ However it is important to check the impact of an alternative approaching H₀ from a specific direction ( at the rate n^1/2). For that, we need tests which are consistent on the whole of H₁. Our idea is to use a marked empirical process based on residuals which converges in distribution to a Gaussian process. To cite this article: M. Harel, C. R. Acad. Sci. Paris, Ser. I 336 (2003).     

Simple estimators for the mean life in destructive attribute life tests

Estimators for the mean life θ of an item whose life length is assumed to be exponentially distributed are considered. The nature of the life test is such that the failure can only be detected at inspection times t₁?t₂? … ?t_m. The inspection destructive, i.e., the items inspected at time t_i are withdrawn from the test regardless of the inspection result.     

Multi-variable weighted geometric means of positive definite matrices

We define a family of weighted geometric means {G(t;ω;A)}_t∈[0,1]ⁿ where ω and A vary over all positive probability vectors in Rⁿ and n-tuples of positive definite matrices resp. Each of these weighted geometric means interpolates between the weighted ALM (t=0_n) and BMP (t=1_n) geometric means (ALM and BMP geometric means have been defined by Ando-Li-Mathias and Bini-Meini-Poloni, respectively.) We show that the weighted geometric means satisfy multidimensional versions of all properties that one would expect for a two-variable weighted geometric mean.     

Time-varying minimum cost flow problems

In this paper, we study a minimum cost flow problem on a time-varying network. Let N(V,A,l,b,c_r,c_w) be a network with an arc set A and a vertex set V. Each a∈A is associated with three integer parameters: a positive transit time b(a,t), an arbitrary transit cost c_r(a,t), and a positive capacity limit l(a,t). Each x∈V is associated with two integer parameters: a waiting cost c_w(x,t) and a vertex capacity l(x,t). All these parameters are functions of the discrete time t=0,1,2,… The objective is to find an optimal schedule to send a flow from the origin (the source vertex) to its destination (the sink vertex) with the minimum cost, subject to the constraint that the flow must arrive at the destination before a deadline T. Three versions of the problem are examined, which are classified depending on whether waiting at the intermediate vertices of the network is strictly prohibited, arbitrarily allowed, or bounded. Three algorithms with pseudopolynomial time complexity are proposed, which can find optimal solutions to the three versions of the problem, respectively.     

The Latent Scale Covariogram: A Tool for Exploring the Spatial Dependence Structure of Nonnormal Responses

When modeling spatially distributed normal responses Y_i in terms of vectors x_i of explanatory variables, one may fit a linear model assuming independence, and then use the empirical variogram of the residuals to determine an appropriate parametric form for the autocorrelation function. Suppose, however, that the responses are not normally distributed—for example, Poisson or Bernoulli. One may model spatial dependence using a hierarchical generalized linear model in which, conditional on a latent Gaussian field Z = {Z_i}, the Y_i have independent distributions from the exponential family, with an appropriate link function connecting their conditional means with the linear predictors x^t_iβ + Z_i. The question then is how to determine an appropriate model for the autocorrelation function of Z. The empirical variogram of the Y_i is no longer appropriate, since (unless the link function is the identity) it is on the wrong scale. We propose here an alternative, the latent scale covariogram, whose graph reflects the autocorrelation structure of the underlying normal field. We illustrate its use on several real datasets, together with a simulated dataset, and obtain results quite different from those obtained using the variogram. Supplementary materials for this article are available online.     

On λ-Fold Equipartite Oberwolfach Problem with Uniform Table Sizes

We consider the following generalization of the Oberwolfach problem:”At a gathering there are n delegations each having m people. Is it possible to arrange a seating of mn people present at s round tables T ₁, T ₂, . . . , T _s (where each T _i can accommodate \( t_i\geq3 \) people and \( \sum t_i = mn \)) for k different meals so that each person has every other person not in the same delegation for a neighbor exactly λ times?” For λ= 1, Liu has obtained the complete solution to the problem when all tables accommodate the same number t of people. In this paper, we give the completesolution to the problem for \( \lambda\geq2 \) when all tables have uniform sizes t.     

Calculating all elements of minimal index in the infinite parametric family of simplest quartic fields

It is a classical problem in algebraic number theory to decide if a number field is monogeneous, that is if it admits power integral bases. It is especially interesting to consider this question in an infinite parametric familiy of number fields. In this paper we consider the infinite parametric family of simplest quartic fields K generated by a root ξ of the polynomial P _t (x) = x ⁴ ? tx ³ ? 6x ² +tx+1, assuming that t > 0, t ≠ 3 and t ² +16 has no odd square factors. In addition to generators of power integral bases we also calculate the minimal index and all elements of minimal index in all fields in this family.     

On HNBUE class after specific age

In this article we introduce new classes of life distributions namely harmonic new better (worse) than used in expectation after specific age t₀ HNBUE¹t₀ (HNWUE¹t₀). The closure properties under various reliability operations such as convolution, mixture, mixing and the homogeneous Poisson shock model of these classes are studied. Furthermore, nonparametric test is proposed to test exponentiality vs. the HNBUE¹t₀ class. The critical values and the powers of this test are calculated to assess the performance of the test. It is shown that the proposed test have high efficiencies for LFR and Weibull distributions. Sets of real data are used as examples to elucidate the use of the proposed test for practical problems.     

Simplification through regression analysis on the dynamic response of plates with arbitrary boundary conditions excited by moving inertia load

Dynamic response of a thin rectangular plate traversed by a moving inertia load with arbitrary boundary condition is investigated through this paper. The inertia effect of mass is considered and relevant formulation is established based on the full-term of acceleration, employing the method of Boundary Characteristic Orthogonal Polynomials, BCOP. To acquire the complete solution of partial differential equations governing on the plate, the Galerkin method is used to separate the temporal function from the spatial one. The problem is formulated in the state space and applying the numerical method of Matrix Exponential the complete solution would be achieved. In the numerical studies, a comprehensive parametric study is performed for both cases of loading when inertia effect is included or neglected. Several mass and aspect ratios for the plate with major types of boundary conditions CCCC, SSSS, CFCF and SFSF are accounted for presenting the results. Dynamic amplification factor against velocity parameter is scrutinized within many graphs alongside with a time history analysis of dynamic deflection for the plate's mid-span. Investigating on the dynamic response concludes to the critical boundary condition upon moving mass. By introducing a conversion factor, the margin of inertia and the critical velocity where happened would be achieved, then through a regression analysis a curve fitting model of polynomials is proposed. Corresponding coefficients testify the goodness of fit for such regression which are reported within tables. Referring to this simplified model of conversion factor pertaining to the specific boundary condition, it would be possible to handle the problem in moving load case without undertaking the complexities arisen from inertia contribution into the formulation. Having derived the factor from simplified model which has been calculated for a specific mass and velocity ratio, then multiplying into the moving load response, the complete solution for moving mass would be achieved.     

Jarque–Bera normality test for the driving Lévy process of a discretely observed univariate SDE

In this paper, we study the Jarque–Bera test for a class of univariate parametric stochastic differential equations (SDE) dX _t  = b(X _t , α)dt + dZ _t , constructed based on the sample observed at discrete time points \({t^{n}_{i}=ih_{n}}\) , i = 1, 2, . . . , n, where Z is a nondegenerate Lévy process with finite moments and h is a sequence of positive real numbers with nh _n → ∞ and \({nh_{n}^{2} \to 0}\) as n → ∞. It is shown that under proper conditions, the Jarque–Bera test statistic based on the Euler residuals can be used to test for the normality of the unobserved Z and the proposed test is consistent against the presence of any nontrivial jump components. Our result indicates that the Jarque–Bera test is easy to implement and asymptotically distribution-free with no fine-tuning parameters. Simulation results to validate the test are given for illustration.     

The Complexity and Effectiveness of Prediction Algorithms

The problem of predicting an arbitrary sequence x₁x₂x₃ · · · is considered with x_{t + 1} being predicted from an analysis of the word x₁x₂ · · · x_t. There are no presumptions concerning the probability structure on this process. The relation between prediction effectiveness and Kolmogorov complexity is established. Then the Hausdorff dimension of sets of sequences for which effective methods of prediction exist is estimated. The prediction method which is asymptotically superior to other arbitrary ones realized by finite automata is constructed.     

Remarks on semilinear partial functional-differential equations with infinite delay

In this paper we will be concerned with the abstract semilinear equation x(Φ)(t) = Ax(Φ)(t) + F(t, x_t(Φ)), t ? τ, x_τ(Φ) = Φ, where A represents the infinitesimal generator of a linear C₀-semigroup in a Banach space and x_～(Φ) denotes the segment of the solution x due to the initial function Φ taken from ?∞ up to t. Conditions for existence, continuous dependence, and regularity of solutions are given which are independent of the topology of the phase space, as well as conditions which guarantee the compactness of bounded orbits, even if exp(At) is not compact for t > 0.