A Base R, simple implementation of the No-Underrun Sampler. This package aims to mostly directly implement the algorithm as described by th
Initial implementation
+25
-5
+1
-1
DESCRIPTION
+1
-1
DESCRIPTION
···
6
6
Description: A bare-bones R implementation of the No-Underrun Sampler
7
7
(Nawaf Bou-Rabee, Bob Carpenter, Sifan Liu, Stefan Oberdörster. 2025.)
8
8
written in base R. This implementation attempts to stick very closely
9
-
to the algorithm described in Appendix A of the paper.
9
+
to the algorithm described in Section 2 and Appendix A of the paper.
10
10
License: MIT + file LICENSE
11
11
Suggests:
12
12
matrixStats
+2
-2
R/NURS.R
+2
-2
R/NURS.R
···
45
45
# hit
46
46
z <- rnorm(d)
47
47
rho <- z / sqrt(sum(z^2))
48
+
# https://artowen.su.domains/mc/Ch-randvectors.pdf (pg. 24)
48
49
49
50
# run
50
51
s <- runif(1, -h / 2, h / 2)
···
55
56
)
56
57
theta + s * rho else theta
57
58
58
-
# helpers
59
59
log_eps_h <- log(epsilon) + log(h)
60
60
orbit_points <- list(theta0)
61
61
log_vals <- logpdf(theta0)
62
62
63
-
# bookkeeping to get ends
63
+
# bookkeeping to get orbit ends
64
64
left <- right <- 1
65
65
66
66
# Orbit selection procedure
+9
-1
README.Rmd
+9
-1
README.Rmd
···
18
18
<!-- badges: start -->
19
19
<!-- badges: end -->
20
20
21
-
Base R implementation of the No-Underrun Sampler.
21
+
A Base R, simple implementation of the No-Underrun Sampler. This implementation aims to mostly directly implement the algorithm as described by the paper, with at most small changes for code aesthetics and performance.
22
22
23
23
## Installation
24
24
···
54
54
geom_point(alpha = 0.3) +
55
55
theme_minimal()
56
56
```
57
+
58
+
## References
59
+
60
+
* Nawaf Bou-Rabee, Bob Carpenter, Sifan Liu, Stefan Oberdörster. 2025.
61
+
[The No-Underrun Sampler: A locally adaptive, gradient free, MCMC
62
+
method](https://arxiv.org/abs/2501.18548v2). *arXiv* 2501.18548 v2.
63
+
64
+
* Art B. Owen. 2013. [Monte Carlo theory, methods and examples](https://artowen.su.domains/mc/). *artowen.su.domains*.
+13
-1
README.md
+13
-1
README.md
···
7
7
8
8
<!-- badges: end -->
9
9
10
-
Base R implementation of the No-Underrun Sampler.
10
+
A Base R, simple implementation of the No-Underrun Sampler. This
11
+
implementation aims to mostly directly implement the algorithm as
12
+
described by the paper, with at most small changes for code aesthetics
13
+
and performance.
11
14
12
15
## Installation
13
16
···
46
49
```
47
50
48
51
<img src="man/figures/README-funnel-1.png" width="100%" />
52
+
53
+
## References
54
+
55
+
- Nawaf Bou-Rabee, Bob Carpenter, Sifan Liu, Stefan Oberdörster. 2025.
56
+
[The No-Underrun Sampler: A locally adaptive, gradient free, MCMC
57
+
method](https://arxiv.org/abs/2501.18548v2). *arXiv* 2501.18548 v2.
58
+
59
+
- Art B. Owen. 2013. [Monte Carlo theory, methods and
60
+
examples](https://artowen.su.domains/mc/). *artowen.su.domains*.