identify.distribution

Identifies the best-fitting theoretical distribution (normal, t, chi-squared, or F) for a given vector (such as test statistics) using maximum likelihood estimation. Comparison is based on AIC (Akaike Information Criterion) and BIC (Bayesian Information Criterion) values, and a Kolmogorov-Smirnov test is performed on the best-fitting distribution.

# S3 method for class 'distribution'
identify(values)

Arguments

values

Numeric vector of test statistic values (e.g., from a Wilcoxon or t-test). NA values are removed internally.

Value

A list with the following elements:

best:

String indicating the best-fitting distribution based on AIC.

comparison:

Data frame with AIC and BIC values for all successfully fitted distributions.

fits:

Named list of fitdist objects for each fitted distribution.

ks_test_best:

Kolmogorov-Smirnov test result for the best-fitting distribution.

Author

Sebastian Gregoricchio

Examples

result <- identify.distribution(values = DEprot::test.toolbox$diff.exp.limma@analyses.result.list$condition_6h.10nM.E2.vs.6h.DMSO$results$statistic)

result$best
#> [1] "norm"
result$comparison
#>   distribution      AIC      BIC
#> 1         norm 130.5920 134.4161
#> 2            t 132.1949 137.9310