Modeling of time-dependent safety performance using anonymized and aggregated smartphone-based dangerous driving event data

Diagnostic accuracy and optimal cutoff points of risk factors is one of the important issues in medical decisions. In order to reassess the cutoff points of markers, longitudinal and time-to-event data of elderly individuals were collected repeatedly through 3 follow-up stages in the Tehran Lipid and Glucose Study. Time-dependent area under the ROC (receiver operating characteristic) curves (AUCs) based on the joint modeling of longitudinal and time-to-event data technique were measured. AUCs were considered to evaluate the discriminative potential of the models. The joint model produced higher AUC values than the Cox model; therefore, accuracy was improved although it is computationally complicated. The results had some differences with the thresholds reported in guidelines due to specificity to the population and/or the means of estimation methods. The estimated cutoff points with regard to sex can be used as a guideline for the Iranian elderly population.

Download Full-text

A joint model of an internal time-dependent covariate and bivariate time-to-event data with an application to muscular dystrophy surveillance, tracking and research network data

10.17077/etd.9zc96njq ◽

2015 ◽

Author(s):

Ke Liu

Keyword(s):

Muscular Dystrophy ◽

Joint Model ◽

Time Dependent ◽

Research Network ◽

Network Data ◽

Event Data ◽

Time To Event ◽

Internal Time ◽

Time To Event Data

Download Full-text

Time-Dependent Safety Performance Assessment of Highway Bridges Considering Mechanical Behavior of Corroded Steel Reinforcement Subject to Carbonation-Induced Corrosion

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.256-259.1515 ◽

2012 ◽

Vol 256-259 ◽

pp. 1515-1519

Author(s):

Jian Xin Peng ◽

Jian Ren Zhang

Keyword(s):

Mechanical Behavior ◽

Concrete Strength ◽

Highway Bridges ◽

Concrete Cover ◽

Safety Performance ◽

Time Dependent ◽

Concrete Durability ◽

Flexural Failure ◽

Bonding Performance ◽

Live Load Model

A structural probabilistic resistance model was developed taking mechanical behavior of corroded reinforcing bars obtained from rapid corrosion specimens and real bridges and bonding performance between concrete and steels into accounts in this paper. Probabilistic model was used to capture randomness and variabilities of materials, geometrical size, corrosive process and corrosion ingredients. A time-dependent reliability model was used to predict probability of structural flexural failure incorporating existing dead load and live load model. It is found form results that probability of flexural failure for corrosive condition of highway bridges is about 18% higher than that predicted for no deterioration condition over next 100 years. It is from results indicated that an improvement in concrete durability specification such as increasing concrete cover and concrete strength can ameliorate the time dependent safety performance of highway bridges.

Download Full-text