Entering edit mode

4.0 years ago

star
▴
350

In the foloww up of my previous question`C: How to calculate probability of a value under Weibull distribution?`

. I have fitted a set of data into Weibull distribution and extract related parameters. Also, I have another set of data that I would like to calculate the p-value for each given value under the distribution.

I calculated the value of cumulative Weibull distribution for each point like below:

```
pweibull (x,scale,shape)
```

I would like to know can I use the value of `pweibull`

directly as p-value or I should calculate like `p-value= 1 - pweibull`

.

by default it reports

`P[X ≤ x]`

; if you want`P[X > x]`

you can add`lower.tail = F`

as parameterMany thanks for your reply. but for finding p-value shall I consider the value of the function itself or do 1- pweibull()?

You're getting the probability that X<=x or the probability that X>x for X following a Weibull distribution with the given parameters. The p-value is the probability of getting a value at least as big as what's observed if the null hypothesis is true. So whether you want to treat P[X>x] as p-value depends on whether you're testing the null hypothesis that X follows a Weibull distribution with the given parameters.

Thanks for your reply.

My data set is a kind of HiC data that shows the interaction between genomic regions. I would like to find the probability of observation which has stronger interaction under the Weibull distribution.

So the hypothesis is like: if x refers to background and X refers to observations.

H0: X<x< p="">

H1:X>x

In this case whether the p-value = 1-pweibull (X,weiull(parameters),lower.tail = F), or I am wrong?