# Confusion about parameter interpretation in Turing documentation
**URL:** https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655
**Category:** Probabilistic Programming
**Tags:** turing
**Created:** [February 4, 2021, 9:47pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655 "2021-02-04T21:47:45Z")
**Posts on this page:** 11
**Page:** 1
### Author:  [@zenkavi](https://discourse.julialang.org/u/zenkavi)
#### Post date: [February 4, 2021, 9:47pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/1 "2021-02-04T21:47:45Z")
In the [Inference of a SDE](https://turing.ml/dev/tutorials/10-bayesiandiffeq/#inference-of-a-stochastic-differential-equation) section og the Turing tutorials the noise added to the states are described in the noise model with the \phi parameters. Is \sigma supposed to capture observation noise? What is its true value?
---
### Author:  [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)
#### Post date: [February 6, 2021, 7:55am UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/2 "2021-02-06T07:55:28Z")
> [@zenkavi](#):
>
> Is σ\sigma supposed to capture observation noise?
Yes.
> [@zenkavi](#):
>
> What is its true value?
```julia
odedata = Array(sol1) + 0.8 * randn(size(Array(sol1)))
```
0.8. That one needs a lot longer of a chain to converge though.
---
### Author:  [@zenkavi](https://discourse.julialang.org/u/zenkavi)
#### Post date: [February 8, 2021, 5:48pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/3 "2021-02-08T17:48:14Z")
Thank you so much for clarifying! How does one get to 0.8 though?
---
### Author:  [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)
#### Post date: [February 8, 2021, 5:50pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/4 "2021-02-08T17:50:29Z")
I set that when generating the data. In reality, 🤷♂️ it’s measurement noise.
---
### Author:  [@zenkavi](https://discourse.julialang.org/u/zenkavi)
#### Post date: [February 8, 2021, 5:57pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/5 "2021-02-08T17:57:32Z")
Ok 😅 thank you!
---
### Author:  [@Fabian](https://discourse.julialang.org/u/Fabian)
#### Post date: [May 12, 2021, 11:29am UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/6 "2021-05-12T11:29:40Z")
I am a little bit confused since in the mentioned SDE problem `data` is used for inference and not `odedata`, see `model = fitlv(data, prob_sde)`. So, in my understanding the data does not contain measurement noise.
I am asking, because I am trying to model a process with diagonal noise as SDE and in addition with measurement noise.
---
### Author:  [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)
#### Post date: [May 12, 2021, 11:45am UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/7 "2021-05-12T11:45:27Z")
Yeah someone pointed out this typo yesterday, and I have a PR open for updating it. It should be `odedata`
---
### Author:  [@Fabian](https://discourse.julialang.org/u/Fabian)
#### Post date: [May 13, 2021, 8:04pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/8 "2021-05-13T20:04:56Z")
Ok, thanks.
But `odedata` is generated with measurement noise only, i.e. no noise model (no SDE) is used?!
[See](https://github.com/TuringLang/TuringTutorials/blob/46123453566cd4499bd5adae6c1841272df5ec96/tutorials/10-bayesian-differential-equations/10_bayesian-differential-equations.jmd#L50).
Is this intentional or am I missing something?
---
### Author:  [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)
#### Post date: [May 13, 2021, 9:41pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/9 "2021-05-13T21:41:43Z")
> [@Fabian](#):
>
> But `odedata` is generated with measurement noise only, i.e. no noise model (no SDE) is used?!
> [See](https://github.com/TuringLang/TuringTutorials/blob/46123453566cd4499bd5adae6c1841272df5ec96/tutorials/10-bayesian-differential-equations/10_bayesian-differential-equations.jmd#L50).
> Is this intentional or am I missing something?
I mean, you can do it with process noise too. What really matters is that it’s one trajectory of data.
---
### Author:  [@Fabian](https://discourse.julialang.org/u/Fabian)
#### Post date: [May 13, 2021, 9:52pm UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/10 "2021-05-13T21:52:36Z")
Sure, but the expected solution of the sampling is then not close to `p = [1.5,1.0,3.0,1.0,0.1,0.1]` but **0** , or? This might be misleading in this set-up.
---
### Author:  [@ChrisRackauckas](https://discourse.julialang.org/u/ChrisRackauckas)
#### Post date: [May 14, 2021, 3:36am UTC](https://discourse.julialang.org/t/confusion-about-parameter-interpretation-in-turing-documentation/54655/11 "2021-05-14T03:36:16Z")
Yeah which is why using one output from the deterministic model (or mean of SDE outputs) is better for testing