Simplified interface that automatically selects appropriate models and
computes per-observation surprise from the model priors.
Usage
auto_surprise(
observed,
expected = NULL,
sample_size = NULL,
include_gaussian = FALSE,
include_sampled = FALSE,
signed = TRUE,
...
)
Arguments
- observed
Numeric vector of observed values
- expected
Numeric vector of expected values (optional)
- sample_size
Numeric vector of sample sizes (optional)
- include_gaussian
Include Gaussian model?
- include_sampled
Include KDE model?
- signed
Compute signed surprise?
- ...
Additional arguments
Value
A bs_surprise object
Examples
observed <- c(50, 100, 150, 200, 75)
expected <- c(10000, 50000, 100000, 25000, 15000)
result <- auto_surprise(observed, expected)