scholarly journals Parsimonious Predictive Mortality Modeling by Regularization and Cross-Validation with and without Covid-Type Effect

Risks ◽  
2020 ◽  
Vol 9 (1) ◽  
pp. 5
Author(s):  
Karim Barigou ◽  
Stéphane Loisel ◽  
Yahia Salhi
Keyword(s):  
Life Insurance ◽  
Cross Validation ◽  
Cohort Effect ◽  
Information Criterion ◽  
Mortality Forecasting ◽  
Hand Model ◽  
Regularization Techniques ◽  
Bayes Information Criterion ◽  
The One ◽  
Regularized Model

Predicting the evolution of mortality rates plays a central role for life insurance and pension funds. Standard single population models typically suffer from two major drawbacks: on the one hand, they use a large number of parameters compared to the sample size and, on the other hand, model choice is still often based on in-sample criterion, such as the Bayes information criterion (BIC), and therefore not on the ability to predict. In this paper, we develop a model based on a decomposition of the mortality surface into a polynomial basis. Then, we show how regularization techniques and cross-validation can be used to obtain a parsimonious and coherent predictive model for mortality forecasting. We analyze how COVID-19-type effects can affect predictions in our approach and in the classical one. In particular, death rates forecasts tend to be more robust compared to models with a cohort effect, and the regularized model outperforms the so-called P-spline model in terms of prediction and stability.

scholarly journals Interclass Interference Suppression in Multi-Class Problems

2021 ◽  
Vol 11 (1) ◽  
pp. 450
Author(s):  
Jinfu Liu ◽  
Mingliang Bai ◽  
Na Jiang ◽  
Ran Cheng ◽  
Xianling Li ◽  
...  
Keyword(s):  
Classification Accuracy ◽  
Cross Validation ◽  
Selection Process ◽  
Interference Suppression ◽  
Generalization Ability ◽  
Suppression Effect ◽  
Binary Classifiers ◽  
The One ◽  
Fold Cross Validation ◽  
Validation Experiments

Multi-classifiers are widely applied in many practical problems. But the features that can significantly discriminate a certain class from others are often deleted in the feature selection process of multi-classifiers, which seriously decreases the generalization ability. This paper refers to this phenomenon as interclass interference in multi-class problems and analyzes its reason in detail. Then, this paper summarizes three interclass interference suppression methods including the method based on all-features, one-class classifiers and binary classifiers and compares their effects on interclass interference via the 10-fold cross-validation experiments in 14 UCI datasets. Experiments show that the method based on binary classifiers can suppress the interclass interference efficiently and obtain the best classification accuracy among the three methods. Further experiments were done to compare the suppression effect of two methods based on binary classifiers including the one-versus-one method and one-versus-all method. Results show that the one-versus-one method can obtain a better suppression effect on interclass interference and obtain better classification accuracy. By proposing the concept of interclass inference and studying its suppression methods, this paper significantly improves the generalization ability of multi-classifiers.

To the problem of uncertainty in interpolation of annual runoff

2012 ◽  
Vol 43 (6) ◽  
pp. 833-850 ◽  
Author(s):  
Ziqi Yan ◽  
Lars Gottschalk ◽  
Irina Krasovskaia ◽  
Jun Xia
Keyword(s):  
Cross Validation ◽  
Drainage Basin ◽  
Annual Runoff ◽  
Mean Value ◽  
Regular Grid ◽  
Main River ◽  
Average Annual Runoff ◽  
The One ◽  
Better Than

The long-term mean value of runoff is the basic descriptor of available water resources. This paper focuses on the accuracy that can be achieved when mapping this variable across space and along main rivers for a given stream gauging network. Three stochastic interpolation schemes for estimating average annual runoff across space are evaluated and compared. Two of the schemes firstly interpolate runoff to a regular grid net and then integrate the grid values along rivers. One of these schemes includes a constraint to account for the lateral water balance along the rivers. The third scheme interpolates runoff directly to points along rivers. A drainage basin in China with 20 gauging sites is used as a test area. In general, all three approaches reproduce the sample discharges along rivers with postdiction errors along main river branches around 10%. Using more objective cross-validation results, it was found that the two schemes based on basin integration, and especially the one with a constraint, performed significantly better than the one with direct interpolation to points along rivers. The analysis did not allow identification of possible influence of surface water use.

Evaluation of Safety Performance Functions Based on Experimental Design Using a Cross-Validation Method

2017 ◽  
Vol 2636 (1) ◽  
pp. 62-72 ◽  
Author(s):  
Jung-Han Wang ◽  
Mohamed A. Abdel-Aty ◽  
Jaeyoung Lee
Keyword(s):  
Goodness Of Fit ◽  
Cross Validation ◽  
Safety Performance ◽  
Major Road ◽  
Highway Safety Manual ◽  
Annual Average Daily Traffic ◽  
Minor Road ◽  
Safety Performance Functions ◽  
The One ◽  
Leave One Out

The Highway Safety Manual (HSM) Part C provides a series of safety performance functions (SPFs) for different roadway conditions. The SPFs suggested in the HSM are formulated on the basis of exposure variables: the logarithms of the annual average daily traffic (AADT) on the major road and on the minor road under the base condition. In this research, data from 7,802 intersections in Florida were collected and processed. These intersections were categorized into seven types based on area type (rural or urban), number of legs (three or four), and number of approaches controlled by stop signs. Twenty-two SPF formulations, including the one suggested by the HSM, were developed for each intersection type for examination of the goodness-of-fit measures of the SPFs. In addition, the goodness of fit of each model of the 22 SPFs in each category was examined with 10-fold leave-one-out cross-validation (LOOCV). With a comparison of the delta values generated with the LOOCV method, it is suggested that the SPF with the logarithm of the total entering vehicle volume and the ratio of the AADT on the minor road and the AADT on the major road are important. In addition, the SPFs with the AADT on the major road and the AADT on the minor road and their logarithmic transformations are also important. Therefore, it is suggested that the future HSM compare these two SPF formulations—as suggested in the current research, along with the original SPF formulation in the manual—and select the one with the best model fit on the basis of the delta value using LOOCV.

scholarly journals Deep permafrost evolution in unstable slopes during the Holocene

2018 ◽  
Author(s):  
Julien Seguinot
Keyword(s):  
Slope Failure ◽  
Time Lag ◽  
Great Influence ◽  
Ground Surface ◽  
Heat Diffusion ◽  
Failure Initiation ◽  
Finite Element Software ◽  
Hand Model ◽  
The One ◽  
Ice Pressure

Large alpine landslides dynamics are generally associated with Quaternary glacier retreat. Some recent datations demonstrate that several thousand years can separate the slope failure initiation from ice pressure unloading. The current study addresses the question whether the persistence of deep permafrost could produce this time lag. A model of deep permafrost evolution is developed, including heat diffusion, phase change and a ground surface transfer function. It is numerically implemented by a 1D finite difference code on the one hand and into a 2D finite element software on the other hand. Model results reveals the great influence of porosity and near-ground processes in permafrost evolution, and illustrates the possible persistence of a permafrost core into the slope.

scholarly journals Tornumonkpe Distribution: Statistical Properties and Goodness of Fit

2021 ◽  
pp. 12-22
Author(s):  
Barinaadaa John Nwikpe
Keyword(s):  
Goodness Of Fit ◽  
Real Life ◽  
Mathematical Expression ◽  
Information Criterion ◽  
Moment Generating Function ◽  
Data Sets ◽  
The Moment ◽  
The One ◽  
Coefficient Of Skewness ◽  
New Distribution

A new sole parameter probability distribution named the Tornumonkpe distribution has been derived in this paper. The new model is a blend of gamma (2,  and gamma(3  distributions. The shape of its density for different values of the parameter has been shown.  The mathematical expression for the moment generating function, the first three raw moments, the second and third moments about the mean, the distribution of order statistics, coefficient of variation and coefficient of skewness has been given. The parameter of the new distribution was estimated using the method of maximum likelihood. The goodness of fit of the Tornumonkpe distribution was established by fitting the distribution to three real life data sets. Using -2lnL, Bayesian Information Criterion (BIC), and Akaike Information Criterion(AIC) as criterial for selecting the best fitting model, it was revealed that the new distribution outperforms the one parameter exponential, Shanker and Amarendra distributions for the data sets used.

The backward problem of parabolic equations with the measurements on a discrete set

2020 ◽  
Vol 28 (1) ◽  
pp. 137-144 ◽  
Author(s):  
Jin Cheng ◽  
Yufei Ke ◽  
Ting Wei
Keyword(s):  
Parabolic Equations ◽  
Numerical Approximation ◽  
Cross Validation ◽  
Continuation Method ◽  
Measurement Data ◽  
Initial Value ◽  
One Dimensional ◽  
Backward Problem ◽  
The One ◽  
Validation Rule

AbstractThe backward problems of parabolic equations are of interest in the study of both mathematics and engineering. In this paper, we consider a backward problem for the one-dimensional heat conduction equation with the measurements on a discrete set. The uniqueness for recovering the initial value is proved by the analytic continuation method. We discretize this inverse problem by a finite element method to deduce a severely ill-conditioned linear system of algebra equations. In order to overcome the ill-posedness, we apply the discrete Tikhonov regularization with the generalized cross validation rule to obtain a stable numerical approximation to the initial value. Numerical results for three examples are provided to show the effect of the measurement data.

Factor Structure of the Italian Version of the Body Satisfaction Scale: A Multisample Analysis

1998 ◽  
Vol 86 (3) ◽  
pp. 1083-1088 ◽  
Author(s):  
Giuseppe Riva ◽  
Enrico Molinari
Keyword(s):  
Body Satisfaction ◽  
Cross Validation ◽  
Clinical Sample ◽  
The Body ◽  
Self Report ◽  
Body Parts ◽  
Factorial Structure ◽  
Italian Version ◽  
Satisfaction Scale ◽  
The One

The paper describes the factorial structure of the Italian version of the Body Satisfaction Scale, a simple self-report questionnaire designed to assess satisfaction with 16 body parts. The results suggest that the structure of the questionnaire can be adequately represented by three different factors. Even if this solution is different from the one obtained in the original English samples, this interpretation was confirmed by a cross-validation on 806 subjects, a clinical sample and two normal samples. This result may be useful for the screening of subjects at risk for eating-disorders because it makes possible prediction of targeted areas of dissatisfaction which is not generally possible with other inventories. The analysis of the factor loadings also suggested that three items should be removed (teeth, eyes, and ears) from the Italian version

scholarly journals INSURANCE RISK WITH VARIABLE NUMBER OF POLICIES

2008 ◽  
Vol 22 (2) ◽  
pp. 213-219
Author(s):  
Ivo Adan ◽  
Vidyadhar Kulkarni
Keyword(s):  
Life Insurance ◽  
Poisson Model ◽  
Variable Number ◽  
Insurance Fund ◽  
Insurance Company ◽  
Insurance Risk ◽  
Constant Rate ◽  
Remarkable Result ◽  
New Policies ◽  
The One

In this article we consider an insurance company selling life insurance policies. New policies are sold at random points in time, and each policy stays active for an exponential amount of time with rate μ, during which the policyholder pays premiums continuously at rate r. When the policy expires, the insurance company pays a claim of random size. The aim is to compute the probability of eventual ruin starting with a given number of policies and a given level of insurance fund. We establish the remarkable result that the ruin probability is identical to the one in the standard compound Poisson model where the insurance fund increases at constant rate r and claims occur according to a Poisson process with rate μ.

scholarly journals Modelling and Prediction of the Gross Mortality Rate in Ecuador

2018 ◽  
Vol 6 (3) ◽  
Author(s):  
Mónica Mite ◽  
Sandra Garcia-Bustos ◽  
Marcela Pincay ◽  
Ana Debón ◽  
Francisco Santoja
Keyword(s):  
Mortality Rate ◽  
Cohort Effect ◽  
Mortality Data ◽  
Arima Models ◽  
Men And Women ◽  
Mortality Forecasting ◽  
Basic Model ◽  
Adjusted Model ◽  
Age Range ◽  
Carter Model

This paper presents the results obtained from the modelling of the mortality data in Ecuador from 1990 to 2010, using the StMoMo library in the open source programming language R. This library was developed based on the Generalized Age-Period-Cohort Models (GAPC), among which is the Lee-Carter model, which has been widely applied in the actuarial area. The gross mortality rate of men and women in an age range of 1 to 85 years was modelled for the data of Ecuador, in the period 1990-2010. Of a total of eight models, two models have been selected because they present a good fit of the data for both genders. The first is the basic model of Lee-Carter and the second, the Plat model, which incorporates the cohort effect. A comparison was made with the two models to determine which one has a better forecast in a horizon of 20 years for specific ages. Both models show and predict the decrease in mortality in Ecuador of both genders, a decrease that is more pronounced, in general, for women at certain ages. In determining the uncertainty of the models, the bootstrap technique was used to define the confidence intervals of the adjusted model. The GAPC and ARIMA models were also compared; the former improve the mortality forecasting.

Sign in / Sign up

Export Citation Format

Share Document