Scribd
Upload
7 views2 pages

Understanding Survival Analysis and Weibull Distribution

This document discusses different ways of representing information contained in duration times using probability distributions. It introduces the density function f(t), cumulative density function F(t), survival function S(t), and hazard function H(t). As an example, it then describes the Weibull distribution for durations, which has scale parameter λ and shape parameter γ. The Weibull hazard, survival function, and density function are defined in terms of λ and γ.

Uploaded by

Ekha
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
7 views2 pages

Understanding Survival Analysis and Weibull Distribution

This document discusses different ways of representing information contained in duration times using probability distributions. It introduces the density function f(t), cumulative density function F(t), survival function S(t), and hazard function H(t). As an example, it then describes the Weibull distribution for durations, which has scale parameter λ and shape parameter γ. The Weibull hazard, survival function, and density function are defined in terms of λ and γ.

Uploaded by

Ekha
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

Suppose event or survival times T are recorded in continuous time.

Then the density

f (t)

of these time defines the probability that an event occurs in the interval

(t , t+ dt) ,

f (t)= lim Pr (t T t+ dt)/dt


dt 0

With cumulative density


t

F ( t )= f ( u ) du
0

From this density the information contained in duration times can be represented in
two different ways. The first involves the chance of surviving until at least time
(or not undergoing the event before duration

t , namely

S (t)=Pr(Tt )=1F (t)


t

f ( u ) du
0

The other way of representing the information involves the hazard rate, measuring
the intensity of the event as a function of time,

H (t)=f (t )/S (t)

(t , t+ dt)

And in probability term, the chance of an event in the interval

given
t

survival until

t . From

h(t ) is obtained the cumulative hazard

and one may also write the survivor function as

H (t )= h ( u ) du ,
0

S (t)=exp(H (t)).

As an example of a parameterized form of time dependence, we may consider the


Weibull distribution for durations

W ( , ) , where

and

are scale and

shape parameters, respectively (Kim and Ibrahim, 2000). The Weibull hazard is
defined as

H (t) t 1
With survival function

S ( t )=exp ( t )
And density

f ( t )= t

exp( t )

The Weibull hazard is monotonically increasing or decreasing in time according to


whether

>1 or

<1 . The value

durations with parameter

=1

leads to exponentially distributed

Peter Congdon.2007. Applied Bayesian [Link] Sussex PO19 8SQ,England

You might also like