Bayesian predictive analysis for Weibull-Pareto composite model with an application to insurance data

Background: We study the impact of suicide-exclusion periods, common in life insurance policies in Australia, on suicide and accidental death rates for life-insured individuals. If a life-insured individual dies by suicide during the period of suicide exclusion, commonly 13 months, the sum insured is not paid. Aims: We examine whether a suicide-exclusion period affects the timing of suicides. We also analyze whether accidental deaths are more prevalent during the suicide-exclusion period as life-insured individuals disguise their death by suicide. We assess the relationship between the insured sum and suicidal death rates. Methods: Crude and age-standardized rates of suicide, accidental death, and overall death, split by duration since the insured first bought their insurance policy, were computed. Results: There were significantly fewer suicides and no significant spike in the number of accidental deaths in the exclusion period for Australian life insurance data. More suicides, however, were detected for the first 2 years after the exclusion period. Higher insured sums are associated with higher rates of suicide. Conclusions: Adverse selection in Australian life insurance is exacerbated by including a suicide-exclusion period. Extension of the suicide-exclusion period to 3 years may prevent some “insurance-induced” suicides – a rationale for this conclusion is given.

Download Full-text

Small Area Health Insurance Estimates Will Fill Gap in Available Health Insurance Data

PsycEXTRA Dataset ◽

10.1037/e559362006-004 ◽

2006 ◽

Cited By ~ 2

Keyword(s):

Health Insurance ◽

Small Area ◽

Health Insurance Data ◽

Insurance Data ◽

Area Health

Download Full-text

How a predictive analysis can be used to assess which hemiparetic patients can gain advantage from additional visual feedback of center-of-pressure displacements?

Movement & Sport Sciences - Science & Motricité ◽

10.1051/sm/2016007 ◽

2016 ◽

pp. 53-57

Author(s):

Bastien Guillebastre ◽

Samir Boudrahem ◽

Patrice Rougier

Keyword(s):

Visual Feedback ◽

Center Of Pressure ◽

Predictive Analysis

Download Full-text

Multilocus Analysis of Genetic Susceptibility to Myocardial Infarction in Russians: Replication Study

Acta Naturae ◽

10.32607/20758251-2017-9-4-74-83 ◽

2017 ◽

Vol 9 (4) ◽

pp. 74-83

Author(s):

N. G. Kukava ◽

B. V. Titov ◽

G. J. Osmak ◽

N. A. Matveeva ◽

O. G. Kulakova ◽

...

Keyword(s):

Myocardial Infarction ◽

Genetic Markers ◽

Prognostic Significance ◽

Cumulative Effect ◽

Mthfr Gene ◽

Composite Model ◽

Replication Study ◽

Epistatic Interactions ◽

Multilocus Analysis ◽

Lipid Metabolism Genes

In search of genetic markers of myocardial infarction (MI) risk, which have prognostic significance for Russians, we performed a replication study of MI association with genetic variants of PCSK9 (rs562556), APOE (epsilon polymorphism, rs7412 and rs429358), LPL (rs320), MTHFR (rs1801133), eNOS (rs2070744), and the 9p21 region (rs1333049) in 405 patients with MI and 198 controls. Significant MI association was observed with variants of the lipid metabolism genes (PCSK9, APOE and LPL), and of eNOS. The SNPs in the MTHFR gene and the 9p21 region were not significantly associated with MI one by one but were included in several different MI-associated allelic combinations identified by multilocus analysis. Since we have not revealed nonlinear epistatic interactions between the components of the identified combinations, we postulate that the cumulative effect of genes that form a combination arises from the summation of their small independent contributions. The prognostic significance of the additive composite model built from the PCSK9, APOE, LPL, and eNOS genes as genetic markers was assessed using ROC analysis. After we included these markers in the previously published composite model of individual genetic risk of MI, the prognostic efficacy in our sample reached AUC = 0.676. However, the results obtained in this study certainly need to be replicated in an independent sample of Russians.

Download Full-text

MONITORING OF OPERATIONAL LOGISTIC PROCESSES IN GENERAL CARGO WAREHOUSES USING PREDICTIVE ANALYSIS

Advanced Logistic Systems - Theory and Practice ◽

10.32971/als.2019.009 ◽

2019 ◽

Vol 13 (1) ◽

pp. 27-36

Author(s):

Andreas Neubert

Keyword(s):

Data Mining ◽

Human Error ◽

Predictive Analysis ◽

Data Mining Algorithm ◽

It Infrastructure ◽

Monitoring Procedure ◽

Manual Recording ◽

Mining Tool ◽

General Cargo ◽

Different Characteristics

Due to the different characteristics of the piece goods (e.g. size and weight), they are transported in general cargo warehouses by manually-operated industrial trucks such as forklifts and pallet trucks. Since manual activities are susceptible to possible human error, errors occur in logistical processes in general cargo warehouses. This leads to incorrect loading, stacking and damage to storage equipment and general cargo. It would be possible to reduce costs arising from errors in logistical processes if these errors could be remedied in advance. This paper presents a monitoring procedure for logistical processes in manually-operated general cargo warehouses. This is where predictive analysis is applied. Seven steps are introduced with a view to integrating predictive analysis into the IT infrastructure of general cargo warehouses. These steps are described in detail. The CRISP4BigData model, the SVM data mining algorithm, the data mining tool R, the programming language C++ for the scoring in general cargo warehouses represent the results of this paper. After having created the system and installed it in general cargo warehouses, initial results obtained with this method over a certain time span will be compared with results obtained without this method through manual recording over the same period.

Download Full-text