protocol

anomaly_dataset_protocol

Protocol for datasets used with anomaly-detection algorithms.

Availability:
logtalk_load(anomaly_detection_protocols(loader))
Author: Paulo Moura
Version: 1:0:0
Date: 2026-04-19
Compilation flags:
static
Dependencies:
(none)
Remarks:
(none)
Inherited public predicates:
(none)

Public predicates

attribute_values/2

Enumerates by backtracking the attributes and their possible values. For discrete attributes, Values is a list of possible values. For continuous attributes, Values is the atom continuous.

Compilation flags:
static
Template:
attribute_values(Attribute,Values)
Mode and number of proofs:
attribute_values(?atom,-list(atom)) - zero_or_more
attribute_values(?atom,-atom) - zero_or_more

class/1

Returns the name of the target label attribute.

Compilation flags:
static
Template:
class(Class)
Mode and number of proofs:
class(-atom) - one

class_values/1

Returns the list of possible labels for the target attribute. Current anomaly benchmarks conventionally use [normal, anomaly].

Compilation flags:
static
Template:
class_values(Values)
Mode and number of proofs:
class_values(-list(atom)) - one

example/3

Enumerates by backtracking the examples in the dataset. Each example has an Id, a Class label, and a list of Attribute-Value pairs.

Compilation flags:
static
Template:
example(Id,Class,AttributeValues)
Mode and number of proofs:
example(-integer,-atom,-list(pair)) - zero_or_more

Protected predicates

(none)

Private predicates

(none)

Operators

(none)