summary.BayesECMpred {ezECM} | R Documentation |
Summary of Unlabeled Event Categorization
Description
Tabulates results from the predict.BayesECM()
function for quick analysis.
Usage
## S3 method for class 'BayesECMpred'
summary(object, index = 1, category = 1, C = 1 - diag(2), ...)
Arguments
object |
an object of |
index |
integer stipulating the event of interest. Value corresponds to the row index of |
category |
integer for the index of the category of interest for hypothesis testing. Alternatively, a character string naming the category of interest can be provided. |
C |
square matrix of dimension 2, providing loss values to be used in hypothesis testing. See Details. |
... |
not used |
Details
Expected Loss
summary.BayesECMpred()
prints expected loss for the binary hypothesis stipulated by category
. Expected loss is calculated using the loss matrix specified with argument C
. The default values for C
result in 0-1 loss being used. Format details for the loss matrix can be found in vignette("syn-data-code")
.
Typicality
Typicality indices are used in cECM_decision()
as part of the decision criteria. Here, we have adapted typicality indices for use with a Bayesian ECM model for outlier detection, when a new observation may not be related to the categories used for training. Probability of the p-value being less than a significance level of 0.05 is reported. If no missing data is used for training, this probability is either 0 or 1.
Value
Prints a summary including probability of each category for the event stipulated by index
, minimum expected loss for binary categorization, and probability of a-typicality of the event for the category specified by category
.
Examples
csv_use <- "good_training.csv"
file_path <- system.file("extdata", csv_use, package = "ezECM")
training_data <- import_pvals(file = file_path, header = TRUE, sep = ",", training = TRUE)
trained_model <- BayesECM(Y = training_data, BT = c(10,1000))
csv_use <- "good_newdata.csv"
file_path <- system.file("extdata", csv_use, package = "ezECM")
new_data <- import_pvals(file = file_path, header = TRUE, sep = ",", training = TRUE)
bayespred <- predict(trained_model, Ytilde = new_data)