Compare k values

Compare the stability and error of various k values to help determine the correct number of signatures (k).

Usage

compare_k_vals(
  musica,
  modality,
  reps = 100,
  min_k = 1,
  max_k = 10,
  error_type = "prop",
  algorithm = "nmf"
)

Arguments

musica

A musica object.

modality

The modality to use, either "SBS96", "DBS78", or "IND83".

reps

Number of times prediction is performed. For each replicate, the count table data is perturbed. Multiple replicates allows for stability analysis by calculating silhouette width on the multiple results. Default 100.

min_k

Lower range of number of signatures for discovery. Default 1.

max_k

Upper range of number of signatures for discovery. Default 10.

error_type

Whether to calculate reconstruction error by proportions ("prop") or raw counts ("raw"). Default "prop".

algorithm

Algorithm for signature discovery. Default "nmf".

Value

a data.frame with stats for each k value tested

Examples

data(musica)
compare_k_vals(musica, "SBS96", reps = 3, min_k = 1, max_k = 5)
#> 
#> k = 1:
#> Calculating signatures on original count table...
#> 
#> Replicate 1 of 3...
#> Replicate 2 of 3...
#> Replicate 3 of 3...
#> 
#> k = 2:
#> Calculating signatures on original count table...
#> 
#> Replicate 1 of 3...
#> Replicate 2 of 3...
#> Replicate 3 of 3...
#> 
#> k = 3:
#> Calculating signatures on original count table...
#> 
#> Replicate 1 of 3...
#> Replicate 2 of 3...
#> Replicate 3 of 3...
#> 
#> k = 4:
#> Calculating signatures on original count table...
#> 
#> Replicate 1 of 3...
#> Replicate 2 of 3...
#> Replicate 3 of 3...
#> 
#> k = 5:
#> Calculating signatures on original count table...
#> 
#> Replicate 1 of 3...
#> Replicate 2 of 3...
#> Replicate 3 of 3...
#> Warning: Removed 1 row containing missing values or values outside the scale range
#> (`geom_line()`).
#> Warning: Removed 1 row containing missing values or values outside the scale range
#> (`geom_point()`).

#>   k  sil_width min_sil_width max_sil_width sd_sil_width sterror_sil_width
#> 1 1         NA            NA            NA           NA                NA
#> 2 2 0.76434462    0.73576651     0.8038029    0.0210121       0.007428898
#> 3 3 0.43603280   -0.07114908     0.8078214    0.2949283       0.085138456
#> 4 4 0.09268357   -0.27594425     0.8124093    0.4133481       0.103337018
#> 5 5 0.06672168   -0.31252943     0.8107605    0.3758637       0.084045671
#>      error min_error max_error  sd_error sterror_error
#> 1 7.009688  6.430394  7.412877 0.4139008     0.2069504
#> 2 5.096309  4.654886  5.443521 0.3370649     0.1685324
#> 3 4.395687  4.031071  4.682656 0.2697601     0.1348800
#> 4 3.379200  3.217509  3.729880 0.2376476     0.1188238
#> 5 2.478940  2.305402  2.793636 0.2163521     0.1081761