Comments (6)
I believe it's only in reglib.py
, but it includes all the rpy2
imports, lines 9-16, as well as what's done in the run_lrgs()
function. So we actually don't need anything in lines 1-75 except for the numpy
and linmix
imports.
from clustr.
okay great, we just put that straight into the clustr.py code. We are having trouble figuring out the run_mcmc part of that function and the .chain part. I haven't seen that before and I'm having trouble relating what I'm finding on the internet to what's actually going on in the code. From what I'm understanding run_mcmc runs linmix through Nmin = 5000 to Nmax = 10000. What do Nmin and Nmax represent and how does that apply to linmix? or should we not worry about it?
from clustr.
I explained this a bit in #33. In principle we would try out every possible combination of values for (slope, intercept, intrinsic_scatter) and use the combination that gives the 'best fit' to the data. But this is quickly a computational nightmare as the time it takes to check all combination grows as the number of grid points ^ # of parameters. So instead there are smarter ways of sampling the parameter space of (slope, intercept, intrinsic_scatter), one of which is a Monte-Carlo Markov Chain (MCMC). It does a random walk in parameter space in a weighted way that will lead to it sampling better fit values than bad fit values. We can talk about this more in a zoom meeting if you'd like.
from clustr.
But it doesn't give a very reliable estimate if you only sample a few points. So you want to run until it 'converges', which is actually a bit of a complicated subject. Setting a minimum number of samples makes sure it has found the "correct" part of parameter space to sample from and the max number of samples will force it to quit if it's struggling to do the fit after a long time.
from clustr.
I think that makes sense, but just to clarify it will run until at least 5000 samples (Nmin) to make sure its reliable, and it will quit after 10000 (Nmax) if it's struggling? A zoom meeting would be helpful. I'll send an email, thanks!
from clustr.
Yes, that's correct. Those values aren't necessarily reliable however - it's always best to inspect the resulting chains by eye. We will do that eventually.
from clustr.
Related Issues (20)
- Fix incorrect scatter plot legend
- Make PEP 8 compatible
- Don't hard-code flags HOT 1
- Switch flag type to enum HOT 1
- Use warnings module
- x axis wrong? HOT 8
- CluStR rewrite!
- Rewrite branch needs to consolidate code HOT 6
- fitter class HOT 6
- using parts other peoples code? HOT 4
- plotting HOT 4
- Complete first-pass run of updated pipeline HOT 12
- run_options HOT 1
- Keyerror Raised HOT 2
- fit function of fitter class won't run HOT 30
- Plotlib.py Rework HOT 7
- Flags for Rewrite HOT 7
- SNR flag HOT 5
- to do HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from clustr.