Carlos Parada

Weighted Arrays?

2

I've been considering this for a while. Would it make sense to define a new struct, a `weighted_array`, which contains both an array and a set of weights? The primary...

Finite Population Corrections for `sem`

1

When sampling from a finite population without replacement, the standard error is less than the current `sem` -- for instance, the standard error is 0 when you've sampled the whole...

Split `AnalyticWeights` into several types

17

As mentioned when working on #754, it might make sense to split `AnalyticWeights` into two separate types of weights: 1. `PrecisionWeights`, which represent observations with a known variance. The weight...

Allow more choices than `true/false` for `corrected`

3

Currently, `corrected` is a `true/false` variable that allows using `n-1` or `n` in the denominator. However, it's possible to use other denominators to get a shrinkage estimator, which can sometimes...

Throw an error for non-integer `FrequencyWeights`

2

Currently, it is possible to construct a set of `FrequencyWeights` with non-integer weights: ``` julia> FrequencyWeights([.5, .1]) 2-element FrequencyWeights{Float64, Float64, Vector{Float64}}: 0.5 0.1 ``` However, the definition of frequency weights...

Array Weights

5

Currently, weights can only be vectors, but there's no reason to rule out array-variate weights. I came across this problem while trying to generalize MCMCChains.jl to allow for weighted samples,...

Questions about `Weights`

10

I have two main questions about the current implementation of weights: 1. Why are they mutable? 2. Why are S and T of different type? It feels very strange to...

Request: L-moments

2

[L-moments](https://en.wikipedia.org/wiki/L-moment) are a set of statistics that can be used in place of regular moments. These statistics are typically more robust against heavy-tailed data, as they exist whenever the mean...

Way to Switch Between LoopVectorization+TensorOperations Dynamically

6

Correct me if I'm wrong, but I believe that LoopVectorization does better than TensorOps for small matrices, while `@tensor` outperforms for large matrices because it calls BLAS. Would it maybe...

@Turbo failing

11

For some reason, `@turbo` throws a warning here at [this specific line in my code](https://github.com/TuringLang/ParetoSmooth.jl/blob/main/src/LeaveOneOut.jl#L203). (Note that this was failing before I made a PR that set `avx=false` to silence...