Skip to content

Commit

Permalink
add the results of some chunks in README and vignettes
Browse files Browse the repository at this point in the history
  • Loading branch information
@MatsuuraKentaro
MatsuuraKentaro committed Oct 12, 2024
1 parent 0badc0f commit 38eb455
Show file tree
Hide file tree
Showing 3 changed files with 85 additions and 0 deletions.
28 changes: 28 additions & 0 deletions README.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -81,6 +81,16 @@ allocation_rule <- learn_allocation_rule(
)
allocation_rule
#> <AllocationRule>
#> dir: allocation_rules/20241012_065853
#> created at: 2024-10-12 10:18:51
#> call:
#> learn_allocation_rule(models = models, N_total = 150, N_ini = rep(10,
#> 5), N_block = 10, Delta = 1.3, outcome_type = "continuous",
#> sd_normal = sqrt(4.5), seed = 123, rl_config = rl_config_set(iter = 1000),
#> alpha = 0.025)
#> iterations: 1000
#> checkpoints: 500, 600, 700, 800, 900, 1000
```

With the default settings, it takes roughly 10-50 seconds per iter, so it would take about 3-14 hours when `iter = 1000`.
Expand All @@ -94,6 +104,16 @@ some_doses <- c( 0, 0, 0, 0, 2, 2, 4, 4, 4, 6, 6, 8, 8, 8)
some_resps <- c(.2, .1, .0, .3, .2, .4, .1, .6, .8, .5, .8, 1.1, .9, 1.6)
allocation_rule$opt_allocation_probs(some_doses, some_resps)
#> 0 2 4 6 8
#> 3.506682e-01 5.533118e-08 6.740401e-02 2.543487e-05 5.819023e-01
```

When 10 subjects in the next block are allocated to each dose according to these probabilities, we recommend using `DoseFinding::rndDesign()`.

```{r eval=FALSE}
probs <- allocation_rule$opt_allocation_probs(some_doses, some_resps)
DoseFinding::rndDesign(probs, 10)
#> [1] 4 0 1 0 5
```

## Simulation-based adjustment of the significance level
Expand All @@ -108,6 +128,7 @@ adjusted_alpha <- adjust_significance_level(
)
adjusted_alpha
#> [1] 0.0221804
```

## How to evaluate the allocation rule
Expand Down Expand Up @@ -140,6 +161,13 @@ for (true_model_name in names(true_response_list)) {
}
head(d_res)
#> simID true_model_name min_p_value selected_model_name estimated_target_dose MAE
#> 1 1 linear 3.637147e-04 linear 6.219219 0.01389732
#> 2 2 linear 1.788856e-03 linear 6.523529 0.03485676
#> 3 3 linear 2.992331e-07 linear 5.812620 0.08700644
#> 4 4 linear 3.650607e-03 linear 8.392319 0.25673228
#> 5 5 linear 2.845267e-04 linear 6.240667 0.01030534
#> 6 6 linear 3.881463e-03 linear 8.081026 0.22689673
```

# Tips
Expand Down
29 changes: 29 additions & 0 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -76,6 +76,16 @@ allocation_rule <- learn_allocation_rule(
)

allocation_rule
#> <AllocationRule>
#> dir: allocation_rules/20241012_065853
#> created at: 2024-10-12 10:18:51
#> call:
#> learn_allocation_rule(models = models, N_total = 150, N_ini = rep(10,
#> 5), N_block = 10, Delta = 1.3, outcome_type = "continuous",
#> sd_normal = sqrt(4.5), seed = 123, rl_config = rl_config_set(iter = 1000),
#> alpha = 0.025)
#> iterations: 1000
#> checkpoints: 500, 600, 700, 800, 900, 1000
```

With the default settings, it takes roughly 10-50 seconds per iter, so
Expand All @@ -91,6 +101,17 @@ some_doses <- c( 0, 0, 0, 0, 2, 2, 4, 4, 4, 6, 6, 8, 8, 8)
some_resps <- c(.2, .1, .0, .3, .2, .4, .1, .6, .8, .5, .8, 1.1, .9, 1.6)

allocation_rule$opt_allocation_probs(some_doses, some_resps)
#> 0 2 4 6 8
#> 3.506682e-01 5.533118e-08 6.740401e-02 2.543487e-05 5.819023e-01
```

When 10 subjects in the next block are allocated to each dose according
to these probabilities, we recommend using `DoseFinding::rndDesign()`.

``` r
probs <- allocation_rule$opt_allocation_probs(some_doses, some_resps)
DoseFinding::rndDesign(probs, 10)
#> [1] 4 0 1 0 5
```

## Simulation-based adjustment of the significance level
Expand All @@ -108,6 +129,7 @@ adjusted_alpha <- adjust_significance_level(
)

adjusted_alpha
#> [1] 0.0221804
```

## How to evaluate the allocation rule
Expand Down Expand Up @@ -144,6 +166,13 @@ for (true_model_name in names(true_response_list)) {
}

head(d_res)
#> simID true_model_name min_p_value selected_model_name estimated_target_dose MAE
#> 1 1 linear 3.637147e-04 linear 6.219219 0.01389732
#> 2 2 linear 1.788856e-03 linear 6.523529 0.03485676
#> 3 3 linear 2.992331e-07 linear 5.812620 0.08700644
#> 4 4 linear 3.650607e-03 linear 8.392319 0.25673228
#> 5 5 linear 2.845267e-04 linear 6.240667 0.01030534
#> 6 6 linear 3.881463e-03 linear 8.081026 0.22689673
```

# Tips
Expand Down
28 changes: 28 additions & 0 deletions vignettes/RLoptimal.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -85,6 +85,16 @@ allocation_rule <- learn_allocation_rule(
)
allocation_rule
#> <AllocationRule>
#> dir: allocation_rules/20241012_065853
#> created at: 2024-10-12 10:18:51
#> call:
#> learn_allocation_rule(models = models, N_total = 150, N_ini = rep(10,
#> 5), N_block = 10, Delta = 1.3, outcome_type = "continuous",
#> sd_normal = sqrt(4.5), seed = 123, rl_config = rl_config_set(iter = 1000),
#> alpha = 0.025)
#> iterations: 1000
#> checkpoints: 500, 600, 700, 800, 900, 1000
```

With the default settings, it takes roughly 10-50 seconds per iter, so it would take about 3-14 hours when `iter = 1000`.
Expand All @@ -98,6 +108,16 @@ some_doses <- c( 0, 0, 0, 0, 2, 2, 4, 4, 4, 6, 6, 8, 8, 8)
some_resps <- c(.2, .1, .0, .3, .2, .4, .1, .6, .8, .5, .8, 1.1, .9, 1.6)
allocation_rule$opt_allocation_probs(some_doses, some_resps)
#> 0 2 4 6 8
#> 3.506682e-01 5.533118e-08 6.740401e-02 2.543487e-05 5.819023e-01
```

When 10 subjects in the next block are allocated to each dose according to these probabilities, we recommend using `DoseFinding::rndDesign()`.

```{r eval=FALSE}
probs <- allocation_rule$opt_allocation_probs(some_doses, some_resps)
DoseFinding::rndDesign(probs, 10)
#> [1] 4 0 1 0 5
```

## Simulation-based adjustment of the significance level
Expand All @@ -112,6 +132,7 @@ adjusted_alpha <- adjust_significance_level(
)
adjusted_alpha
#> [1] 0.0221804
```

## How to evaluate the allocation rule
Expand Down Expand Up @@ -144,6 +165,13 @@ for (true_model_name in names(true_response_list)) {
}
head(d_res)
#> simID true_model_name min_p_value selected_model_name estimated_target_dose MAE
#> 1 1 linear 3.637147e-04 linear 6.219219 0.01389732
#> 2 2 linear 1.788856e-03 linear 6.523529 0.03485676
#> 3 3 linear 2.992331e-07 linear 5.812620 0.08700644
#> 4 4 linear 3.650607e-03 linear 8.392319 0.25673228
#> 5 5 linear 2.845267e-04 linear 6.240667 0.01030534
#> 6 6 linear 3.881463e-03 linear 8.081026 0.22689673
```

# Tips
Expand Down

0 comments on commit 38eb455

Please sign in to comment.