-
Here are my thoughts on a vignette for Gibbs sampling.
We should think about what to assume people already know. For example I think we can assume they know what a conditional distribution is and wha…
-
The [HDP](http://www.cs.berkeley.edu/~jordan/papers/hdp.pdf) paper contains some discussion (below) of posterior sampling the Dirichlet process concentration parameters (`\alpha_0`/`\gamma` or `\dish_…
-
convert Gibbs sampling to C++
-
I have a question re the implementation of the transform method for LDA class. Does using an LDA object and calling `transform` only run the fitted model on the new data or does it fit it again before…
-
Tool for inferring mixture of CNV subclones from ACNV output.
- [x] Develop resources for simulating tumor phylogenies/mixtures
- Wrote python code for simulating phylogenies and generating correspo…
-
In R, you can add spline bases to the design matrix for a linear model with `y ~ x1 + bs(x2)` to model the variable `x2` as a cubic spline with B-spline basis functions. I'd like to implement somethi…
-
E.g. blow will give error.
```julia
d1 = Dual(1.0, 0)
d2 = Dual(1.0, 0, 0)
d1 + d2
```
from
```
error("npartials($(typeof(a))) != npartials($(typeof(b)))")
```
NOTE: ForwardDiff write ru…
-
All SFI solvers include a normalization step before returning the eliminated factors. As far as we can tell, this is unnecessary and can be removed. If we decide to leave it, we should document (in th…
-
bootstraps.gz is being written to the directory from where salmon command is being run with salmon v0.8. It was being written to the correct directory with salmon v0.7. This happens with both bootstra…
-
There might be an error in the GibbsAsk implementation. Evidence for child nodes is ignored. RejectionSampling and Likelyhoodweighting work as expected. Test follows:
// create two nodes: p…