category

dimension_reducer_common

Shared predicates for dimension reducer learning defaults, dataset helpers, transformation, export, and printing.

Availability:
logtalk_load(dimension_reduction_protocols(loader))
Author: Paulo Moura
Version: 1:0:0
Date: 2026-05-07
Compilation flags:
static
Implements:
public dimension_reducer_protocol
Extends:
public options
Uses:
format
linear_algebra
list
numberlist
population
type
user
Remarks:
(none)
Inherited public predicates:
 check_dimension_reducer/1  check_option/1  check_options/1  default_option/1  default_options/1  diagnostic/2  diagnostics/2  dimension_reducer_options/2  export_to_clauses/4  export_to_file/4  learn/2  learn/3  option/2  option/3  print_dimension_reducer/1  transform/3  valid_dimension_reducer/1  valid_option/1  valid_options/1

Public predicates

(no local declarations; see entity ancestors if any)

Protected predicates

check_component_count/3

Checks that a requested component count does not exceed the supported maximum and returns the accepted count.

Compilation flags:
static
Template:
check_component_count(RequestedComponentCount,MaxComponentCount,ComponentCount)
Mode and number of proofs:
check_component_count(+integer,+integer,-integer) - one

dimension_reducer_data/3

Default hook predicate for exposing the learned encoders and projection components from a reducer term. Importing implementations may override it when using a non-standard reducer representation.

Compilation flags:
static
Template:
dimension_reducer_data(DimensionReducer,Encoders,Components)
Mode and number of proofs:
dimension_reducer_data(+compound,-list,-list) - one

dimension_reducer_diagnostics_data/2

Default hook predicate for exposing diagnostics metadata from a reducer term. Importing implementations may override it when using a non-standard reducer representation.

Compilation flags:
static
Template:
dimension_reducer_diagnostics_data(DimensionReducer,Diagnostics)
Mode and number of proofs:
dimension_reducer_diagnostics_data(+compound,-list(compound)) - one

example_attribute_values/2

Hook predicate that importing dimension reducer implementations must define in order to extract example attribute values from their local training example representation.

Compilation flags:
static
Template:
example_attribute_values(Example,AttributeValues)
Mode and number of proofs:
example_attribute_values(+compound,-list(pair)) - one

dataset_attributes/2

Collects the dataset attribute declarations as Attribute-Values pairs.

Compilation flags:
static
Template:
dataset_attributes(Dataset,Attributes)
Mode and number of proofs:
dataset_attributes(+object_identifier,-list(pair)) - one

check_continuous_attributes/1

Checks that all declared dataset attributes are continuous.

Compilation flags:
static
Template:
check_continuous_attributes(Attributes)
Mode and number of proofs:
check_continuous_attributes(+list(pair)) - one

check_examples_non_empty/2

Checks that a training example collection is not empty.

Compilation flags:
static
Template:
check_examples_non_empty(Dataset,Examples)
Mode and number of proofs:
check_examples_non_empty(+object_identifier,+list) - one

check_example_values/2

Checks that all example attribute values are present and numeric for the declared attributes.

Compilation flags:
static
Template:
check_example_values(Examples,AttributeNames)
Mode and number of proofs:
check_example_values(+list,+list(atom)) - one

check_example_attributes/2

Checks that a single example contains numeric values for all declared attributes.

Compilation flags:
static
Template:
check_example_attributes(AttributeNames,AttributeValues)
Mode and number of proofs:
check_example_attributes(+list(atom),+list(pair)) - one

attribute_value/3

Looks up an attribute value in a list of Attribute-Value pairs.

Compilation flags:
static
Template:
attribute_value(Attribute,AttributeValues,Value)
Mode and number of proofs:
attribute_value(+atom,+list(pair),-term) - one

build_encoders/4

Builds continuous feature encoders by computing per-attribute centering and optional scaling statistics.

Compilation flags:
static
Template:
build_encoders(AttributeNames,Examples,Options,Encoders)
Mode and number of proofs:
build_encoders(+list(atom),+list,+list(compound),-list(compound)) - one

base_dimension_reducer_diagnostics/6

Builds common diagnostics metadata terms for a learned reducer and appends reducer-specific diagnostics terms.

Compilation flags:
static
Template:
base_dimension_reducer_diagnostics(Model,AttributeNames,Components,Options,ExtraDiagnostics,Diagnostics)
Mode and number of proofs:
base_dimension_reducer_diagnostics(+atom,+list(atom),+list,+list(compound),+list(compound),-list(compound)) - one

preprocessing_diagnostics/3

Builds shared preprocessing diagnostics metadata from an explicit centering flag and the effective training options.

Compilation flags:
static
Template:
preprocessing_diagnostics(Center,Options,Preprocessing)
Mode and number of proofs:
preprocessing_diagnostics(+boolean,+list(compound),-list(compound)) - one

iterative_dimension_reducer_diagnostics/11

Builds diagnostics metadata for reducers that report sample counts, iterative convergence terms, and optional leading or trailing reducer-specific diagnostics.

Compilation flags:
static
Template:
iterative_dimension_reducer_diagnostics(Model,AttributeNames,Components,SampleCount,Options,LeadingDiagnostics,Convergence,Iterations,FinalDelta,TrailingDiagnostics,Diagnostics)
Mode and number of proofs:
iterative_dimension_reducer_diagnostics(+atom,+list(atom),+list,+integer,+list(compound),+list(compound),+term,+term,+term,+list(compound),-list(compound)) - one

component_iteration_diagnostics/4

Extracts per-component convergence, iteration, and final-delta lists from component diagnostics records.

Compilation flags:
static
Template:
component_iteration_diagnostics(ComponentDiagnostics,Convergences,IterationCounts,FinalDeltas)
Mode and number of proofs:
component_iteration_diagnostics(+list(compound),-list(atom),-list(integer),-list(number)) - one

zero_vector_like/2

Constructs a zero vector matching the length of the first vector in a list of vectors, or returns the empty list when the input is empty.

Compilation flags:
static
Template:
zero_vector_like(Vectors,ZeroVector)
Mode and number of proofs:
zero_vector_like(+list(list(number)),-list(number)) - one

basis_vector/3

Constructs a canonical basis vector for the requested size and one-based index.

Compilation flags:
static
Template:
basis_vector(Size,Index,Vector)
Mode and number of proofs:
basis_vector(+integer,+integer,-list(number)) - one

initial_vectors/2

Constructs the default all-ones initial vector followed by canonical basis vectors for the requested size.

Compilation flags:
static
Template:
initial_vectors(Size,Vectors)
Mode and number of proofs:
initial_vectors(+integer,-list(list(number))) - one

basis_initial_vectors/3

Constructs canonical basis vectors from the given one-based index up to the requested size.

Compilation flags:
static
Template:
basis_initial_vectors(Index,Size,Vectors)
Mode and number of proofs:
basis_initial_vectors(+integer,+integer,-list(list(number))) - one

extract_components/5

Extracts leading positive eigen-components from a numeric matrix using the shared symmetric eigensolver until the requested count or the configured tolerance is reached.

Compilation flags:
static
Template:
extract_components(Matrix,Requested,Options,Components,Eigenvalues)
Mode and number of proofs:
extract_components(+list(list(number)),+integer,+list(compound),-list(list(number)),-list(number)) - one

known_attribute_values/3

Collects the known numeric values for a given attribute across the training examples.

Compilation flags:
static
Template:
known_attribute_values(Examples,Attribute,Values)
Mode and number of proofs:
known_attribute_values(+list,+atom,-list(number)) - one

examples_to_rows/3

Encodes a list of training examples into numeric feature rows using the importing reducer example hook and learned encoders.

Compilation flags:
static
Template:
examples_to_rows(Examples,Encoders,Rows)
Mode and number of proofs:
examples_to_rows(+list,+list(compound),-list(list(number))) - one

encode_instance/3

Encodes an instance using the learned continuous attribute encoders.

Compilation flags:
static
Template:
encode_instance(Encoders,AttributeValues,Features)
Mode and number of proofs:
encode_instance(+list(compound),+list(pair),-list(number)) - one

encoder_attribute_names/2

Collects encoder attribute names preserving encoder order.

Compilation flags:
static
Template:
encoder_attribute_names(Encoders,AttributeNames)
Mode and number of proofs:
encoder_attribute_names(+list(compound),-list(atom)) - one

project_components/4

Projects encoded features onto the learned components and returns component_N-Score pairs.

Compilation flags:
static
Template:
project_components(Components,Features,Index,ReducedInstance)
Mode and number of proofs:
project_components(+list(list(number)),+list(number),+integer,-list(pair)) - one

valid_linear_encoders/1

True when a list of encoders only contains valid continuous/3 encoder terms with distinct attributes.

Compilation flags:
static
Template:
valid_linear_encoders(Encoders)
Mode and number of proofs:
valid_linear_encoders(+list(compound)) - zero_or_one

valid_projection_components/2

True when projection components are numeric vectors compatible with the encoder feature dimension.

Compilation flags:
static
Template:
valid_projection_components(Encoders,Components)
Mode and number of proofs:
valid_projection_components(+list(compound),+list(list(number))) - zero_or_one

valid_dimension_reducer_metadata/1

True when diagnostics metadata records the reducer model and effective training options.

Compilation flags:
static
Template:
valid_dimension_reducer_metadata(Diagnostics)
Mode and number of proofs:
valid_dimension_reducer_metadata(+list(compound)) - zero_or_one

Private predicates

(no local declarations; see entity ancestors if any)

Operators

(none)