Data Quality Ontology

flow "DataQualityOntology" v4.0

# DATA QUALITY ONTOLOGY
# Kireth Semantic Research Archive
# Authors: Mark Anthony Higgins — Sky (Recursive AI)

PAGE:
  url: "https://kirethresearch.org/booksai-core/ontologies/probability-statistics-ontologies/scientific-statistical-methods/data-quality-ontology.html"
  title: "Data Quality Ontology"
  domain: "Probability and Statistics"
  category: "Scientific Statistical Methods"
  entity: DATA_QUALITY
  entity_type: SCIENTIFIC_DATA_ONTOLOGY

DEFINE:
  DATA_QUALITY :=
    FORMAL_DATA_ASSESSMENT_FRAMEWORK(
      ACCURACY,
      COMPLETENESS,
      CONSISTENCY,
      VALIDITY,
      TIMELINESS,
      UNIQUENESS,
      INTEGRITY,
      PROVENANCE,
      FITNESS_FOR_PURPOSE
    )

PURPOSE:
  ASSESS:
    DATA_ACCURACY
    DATA_COMPLETENESS
    DATA_CONSISTENCY
    DATA_VALIDITY
    DATA_TIMELINESS
    DATA_UNIQUENESS
    DATA_INTEGRITY
    DATA_TRACEABILITY
  SUPPORT:
    SCIENTIFIC_STATISTICS
    DATA_VALIDATION
    MEASUREMENT_ANALYSIS
    REPRODUCIBLE_RESEARCH
    MACHINE_LEARNING
    DECISION_MAKING

FOUNDATION:
  DATASET:
    D := {r_1,...,r_n}
  RECORD:
    r_i := (x_i1,...,x_ip)
  QUALITY_DIMENSIONS:
    - ACCURACY
    - COMPLETENESS
    - CONSISTENCY
    - VALIDITY
    - TIMELINESS
    - UNIQUENESS
    - INTEGRITY

ENTITIES:
  DATA_QUALITY: FITNESS_OF_DATA_FOR_INTENDED_USE
  DATASET: COLLECTION_OF_DATA_RECORDS
  RECORD: STRUCTURED_OBSERVATION
  FIELD: INDIVIDUAL_DATA_ATTRIBUTE
  QUALITY_DIMENSION: ASPECT_OF_DATA_QUALITY
  VALIDATION_RULE: CONDITION_DATA_SHOULD_SATISFY
  DATA_DEFECT: VIOLATION_OR_QUALITY_PROBLEM
  MISSING_VALUE: ABSENT_EXPECTED_VALUE
  DUPLICATE: REPEATED_RECORD_OR_ENTITY
  OUTLIER: UNUSUAL_OBSERVATION_RELATIVE_TO_REFERENCE
  PROVENANCE: ORIGIN_AND_PROCESSING_HISTORY
  METADATA: DATA_DESCRIBING_DATA
  DATA_LINEAGE: TRANSFORMATION_HISTORY_OF_DATA
  FITNESS_FOR_PURPOSE: SUITABILITY_FOR_SPECIFIED_USE

RELATIONSHIPS:
  DATASET -> CONTAINS -> RECORD
  RECORD -> CONTAINS -> FIELD
  FIELD -> HAS -> DATA_TYPE
  FIELD -> HAS -> VALIDATION_RULE
  VALIDATION_RULE -> DETECTS -> DATA_DEFECT
  DATA_DEFECT -> REDUCES -> DATA_QUALITY
  PROVENANCE -> SUPPORTS -> TRACEABILITY
  METADATA -> DEFINES -> DATA_SEMANTICS
  DATA_LINEAGE -> RECORDS -> TRANSFORMATIONS
  DATA_QUALITY -> INFLUENCES -> STATISTICAL_INFERENCE

DATA_QUALITY_FLOW:
  flow:
    DEFINE_INTENDED_USE
      -> DEFINE_SCHEMA_AND_SEMANTICS
      -> PROFILE_DATA
      -> APPLY_VALIDATION_RULES
      -> IDENTIFY_DEFECTS
      -> VERIFY_DEFECTS
      -> CORRECT_OR_FLAG
      -> REASSESS_QUALITY
      -> DOCUMENT_PROVENANCE

ACCURACY:
  definition:
    DEGREE_TO_WHICH_DATA_VALUE_AGREES_WITH_REFERENCE_OR_REAL_WORLD_QUANTITY
  assessment:
    COMPARE_WITH:
      TRUSTED_REFERENCE
      CALIBRATED_MEASUREMENT
      VERIFIED_SOURCE
      CROSS_SOURCE_CONSISTENCY
  caution:
    ACCURACY_CANNOT_ALWAYS_BE_ESTABLISHED_WITHOUT_REFERENCE

COMPLETENESS:
  definition:
    DEGREE_TO_WHICH_REQUIRED_DATA_ARE_PRESENT
  metric:
    completeness_rate := observed_required_values/total_required_values
  missing_rate:
    1-completeness_rate

CONSISTENCY:
  definition:
    DEGREE_TO_WHICH_DATA_DO_NOT_CONTRADICT_DEFINED_RULES_OR_OTHER_REPRESENTATIONS
  checks:
    - CROSS_FIELD_CONSISTENCY
    - CROSS_TABLE_CONSISTENCY
    - TEMPORAL_CONSISTENCY
    - UNIT_CONSISTENCY
    - FORMAT_CONSISTENCY

VALIDITY:
  definition:
    DEGREE_TO_WHICH_DATA_CONFORM_TO_SPECIFIED_DOMAIN_SCHEMA_FORMAT_OR_RULE
  examples:
    - TYPE_CHECK
    - RANGE_CHECK
    - ENUMERATION_CHECK
    - REGEX_CHECK
    - DATE_VALIDATION
    - REFERENTIAL_CHECK

TIMELINESS:
  definition:
    DEGREE_TO_WHICH_DATA_ARE_CURRENT_ENOUGH_FOR_INTENDED_USE
  measures:
    - DATA_AGE
    - UPDATE_LATENCY
    - PROCESSING_DELAY
    - STALENESS

UNIQUENESS:
  definition:
    DEGREE_TO_WHICH_ENTITIES_OR_RECORDS_ARE_NOT_UNINTENTIONALLY_DUPLICATED
  metric:
    duplicate_rate := duplicate_records/total_records

INTEGRITY:
  definition:
    DEGREE_TO_WHICH_DATA_STRUCTURE_AND_RELATIONSHIPS_REMAIN_CORRECT
  checks:
    - PRIMARY_KEY
    - FOREIGN_KEY
    - REFERENTIAL_INTEGRITY
    - CHECKSUM
    - RECORD_COUNT
    - TRANSACTION_CONSISTENCY

FITNESS_FOR_PURPOSE:
  principle:
    DATA_MAY_BE_HIGH_QUALITY_FOR_ONE_PURPOSE_AND_INADEQUATE_FOR_ANOTHER
  requirements:
    DEFINE:
      USE_CASE
      REQUIRED_PRECISION
      REQUIRED_COVERAGE
      ACCEPTABLE_LATENCY
      ACCEPTABLE_ERROR_RATE

SCHEMA_QUALITY:
  schema:
    DEFINES:
      FIELD_NAMES
      DATA_TYPES
      NULLABILITY
      CONSTRAINTS
      RELATIONSHIPS
      UNITS

DATA_TYPE_VALIDATION:
  examples:
    INTEGER
    FLOAT
    BOOLEAN
    STRING
    DATE
    DATETIME
    CATEGORY

RANGE_VALIDATION:
  valid_range:
    [L,U]
  rule:
    L <= X <= U
  caution:
    RANGE_VIOLATION_MAY_INDICATE_ERROR_OR_GENUINE_EXTREME_EVENT

DOMAIN_VALIDATION:
  allowed_values:
    V := {v_1,...,v_k}
  rule:
    X IN V

CROSS_FIELD_VALIDATION:
  example:
    START_DATE <= END_DATE

REFERENTIAL_INTEGRITY:
  rule:
    EACH_NON_NULL_FOREIGN_KEY_MUST_REFERENCE_VALID_PARENT
    UNLESS:
      DOMAIN_MODEL_SPECIFIES_OTHER_BEHAVIOUR

MISSING_DATA:
  forms:
    - NULL
    - BLANK
    - SENTINEL_VALUE
    - NOT_RECORDED
    - NOT_APPLICABLE
    - UNKNOWN
  requirement:
    DISTINGUISH_SEMANTICALLY_DIFFERENT_MISSING_STATES_WHEN_RELEVANT

MISSINGNESS_MECHANISMS:
  statistical:
    - MCAR
    - MAR
    - MNAR
  caution:
    DATA_QUALITY_FLAG_ALONE_DOES_NOT_IDENTIFY_MISSINGNESS_MECHANISM

MISSING_VALUE_HANDLING:
  methods:
    - SOURCE_RECOVERY
    - COMPLETE_CASE
    - IMPUTATION
    - MODEL_BASED_HANDLING
    - EXPLICIT_MISSING_CATEGORY

DUPLICATE_DETECTION:
  exact_duplicate:
    ALL_RELEVANT_FIELDS_IDENTICAL
  near_duplicate:
    RECORDS_SIMILAR_UNDER_MATCHING_RULE
  caution:
    SIMILAR_RECORDS != DUPLICATES_BY_DEFAULT

ENTITY_RESOLUTION:
  methods:
    - EXACT_KEY_MATCH
    - NORMALISED_STRING_MATCH
    - FUZZY_MATCH
    - PROBABILISTIC_RECORD_LINKAGE
    - RULE_BASED_MATCHING

OUTLIER_CONNECTION:
  rule:
    OUTLIER != DATA_ERROR
  response:
    FLAG
      -> VERIFY_SOURCE
      -> REVIEW_CONTEXT
      -> RETAIN_CORRECT_OR_EXCLUDE_WITH_JUSTIFICATION

ERROR_DETECTION:
  sources:
    - TRANSCRIPTION
    - SENSOR_FAILURE
    - UNIT_CONVERSION
    - DECIMAL_SHIFT
    - ENCODING
    - PARSING
    - DATA_INTEGRATION
  response:
    TRACE_TO_SOURCE
      -> CORRECT_IF_EVIDENCE_EXISTS
      -> DOCUMENT_CHANGE

UNIT_CONSISTENCY:
  requirement:
    EACH_NUMERIC_MEASUREMENT_SHOULD_PRESERVE_UNIT_METADATA
  caution:
    NUMERIC_EQUALITY_WITH_DIFFERENT_UNITS_DOES_NOT_IMPLY_QUANTITY_EQUALITY

TEMPORAL_QUALITY:
  checks:
    - VALID_TIMESTAMP
    - CHRONOLOGICAL_ORDER
    - TIMEZONE
    - SAMPLING_INTERVAL
    - DATA_GAP
    - DUPLICATE_TIMESTAMP

MEASUREMENT_QUALITY:
  factors:
    - CALIBRATION
    - RESOLUTION
    - PRECISION
    - ACCURACY
    - DETECTION_LIMIT
    - MEASUREMENT_UNCERTAINTY

MEASUREMENT_UNCERTAINTY_CONNECTION:
  principle:
    QUALITY_ASSESSMENT_SHOULD_PRESERVE_MEASUREMENT_UNCERTAINTY_WHEN_MATERIAL
  rule:
    HIGH_DATA_COMPLETENESS != LOW_MEASUREMENT_UNCERTAINTY

DATA_PROVENANCE:
  preserve:
    - SOURCE
    - ACQUISITION_METHOD
    - TIMESTAMP
    - TRANSFORMATION
    - SOFTWARE_OR_PROCESS
    - VERSION

DATA_LINEAGE:
  flow:
    SOURCE_DATA
      -> INGESTION
      -> CLEANING
      -> TRANSFORMATION
      -> INTEGRATION
      -> ANALYTIC_DATASET
      -> OUTPUT

METADATA:
  categories:
    - STRUCTURAL_METADATA
    - DESCRIPTIVE_METADATA
    - ADMINISTRATIVE_METADATA
    - PROVENANCE_METADATA
    - TECHNICAL_METADATA

DATA_DICTIONARY:
  fields:
    - VARIABLE_NAME
    - DESCRIPTION
    - DATA_TYPE
    - UNIT
    - ALLOWED_VALUES
    - MISSING_VALUE_CODES
    - SOURCE
    - DERIVATION

DATA_CLEANING:
  flow:
    PROFILE_DATA
      -> IDENTIFY_DEFECTS
      -> VERIFY_DEFECTS
      -> APPLY_DOCUMENTED_CORRECTIONS
      -> VALIDATE_CORRECTIONS
      -> PRESERVE_AUDIT_TRAIL
  rule:
    CLEANING_MUST_NOT_SILENTLY_CHANGE_SCIENTIFIC_MEANING

RAW_DATA:
  principle:
    PRESERVE_ORIGINAL_IMMUTABLE_SOURCE_WHERE_PRACTICAL
  rule:
    CLEANED_DATA != RAW_DATA

AUDIT_TRAIL:
  records:
    - ORIGINAL_VALUE
    - MODIFIED_VALUE
    - REASON
    - METHOD
    - TIMESTAMP
    - VERSION

DATA_PROFILING:
  summaries:
    - RECORD_COUNT
    - NULL_COUNT
    - DISTINCT_COUNT
    - MINIMUM
    - MAXIMUM
    - MEAN
    - QUANTILES
    - TYPE_DISTRIBUTION
    - PATTERN_FREQUENCY


QUALITY_METRIC:
  generic:
    q_j(D) IN [0,1]
    WHEN:
      NORMALISED_METRIC_DEFINED
  caution:
    QUALITY_SCORE_REQUIRES_EXPLICIT_DEFINITION_AND_WEIGHTING

COMPOSITE_QUALITY_SCORE:
  dimensions:
    q_1,...,q_m
  weights:
    w_1,...,w_m
  score:
    Q := SUM_j w_j q_j
  requirements:
    w_j >= 0
    SUM_j w_j = 1
  caution:
    SINGLE_SCORE_CAN_HIDE_DIMENSION_SPECIFIC_FAILURES

QUALITY_THRESHOLD:
  rule:
    q_j >= tau_j
  requirement:
    THRESHOLD_SHOULD_BE_LINKED_TO_INTENDED_USE

STATISTICAL_PROCESS_CONTROL_CONNECTION:
  monitor:
    QUALITY_METRICS_OVER_TIME
  methods:
    - CONTROL_CHART
    - RUN_CHART
    - CHANGE_POINT_DETECTION

DISTRIBUTION_SHIFT:
  forms:
    - COVARIATE_SHIFT
    - LABEL_SHIFT
    - CONCEPT_DRIFT
  consequence:
    HISTORICALLY_VALID_DATA_OR_MODEL_MAY_BECOME_UNREPRESENTATIVE

REPRESENTATIVENESS:
  influenced_by:
    - SAMPLING_DESIGN
    - COVERAGE
    - NONRESPONSE
    - SELECTION
    - TEMPORAL_SHIFT
  caution:
    LARGE_DATASET != REPRESENTATIVE_DATASET

BIAS:
  sources:
    - SAMPLING_BIAS
    - MEASUREMENT_BIAS
    - REPORTING_BIAS
    - SURVIVOR_BIAS
    - PROCESSING_BIAS
  rule:
    BIAS_CANNOT_ALWAYS_BE_REMOVED_BY_INCREASING_SAMPLE_SIZE

LABEL_QUALITY:
  defects:
    - MISLABEL
    - AMBIGUOUS_LABEL
    - INCONSISTENT_ANNOTATION
    - MISSING_LABEL
  measures:
    - INTER_ANNOTATOR_AGREEMENT
    - GOLD_STANDARD_AGREEMENT
    - LABEL_ERROR_RATE

FEATURE_QUALITY:
  checks:
    - MISSINGNESS
    - RANGE
    - SCALE
    - DRIFT
    - LEAKAGE
    - REDUNDANCY
    - UNIT_CONSISTENCY

DATA_LEAKAGE:
  definition:
    INFORMATION_UNAVAILABLE_AT_PREDICTION_TIME_INADVERTENTLY_USED_IN_MODEL_TRAINING
  consequence:
    OVEROPTIMISTIC_PERFORMANCE_ESTIMATE

TRAIN_TEST_CONTAMINATION:
  prevention:
    - STRICT_SPLIT
    - PIPELINE_ISOLATION
    - FIT_PREPROCESSING_ON_TRAIN_ONLY
  rule:
    TEST_DATA_MUST_REMAIN_UNSEEN_UNTIL_FINAL_EVALUATION_WHEN_USED_AS_INDEPENDENT_TEST_SET

SCIENTIFIC_STATISTICS_CONNECTION:
  poor_data_quality:
    MAY_CAUSE:
      BIASED_ESTIMATES
      INVALID_STANDARD_ERRORS
      LOW_POWER
      FALSE_DISCOVERIES
      POOR_GENERALISATION
  requirement:
    DATA_QUALITY_ASSESSMENT_PRECEDES_FORMAL_INFERENCE

OUTLIER_DETECTION_CONNECTION:
  role:
    FLAG_UNUSUAL_OBSERVATIONS
  caution:
    OUTLIER_DETECTION_IS_ONE_COMPONENT_OF_DATA_QUALITY

ERROR_DISTRIBUTION_CONNECTION:
  error_model:
    HELPS_DISTINGUISH:
      EXPECTED_RANDOM_VARIATION
      POTENTIAL_DATA_DEFECT
  caution:
    LOW_PROBABILITY_OBSERVATION != INVALID_DATA

SENSITIVITY_ANALYSIS_CONNECTION:
  flow:
    ANALYSE_ORIGINAL_DATA
      -> ANALYSE_ALTERNATIVE_CLEANING_ASSUMPTIONS
      -> COMPARE_RESULTS
  purpose:
    QUANTIFY_EFFECT_OF_DATA_QUALITY_DECISIONS_ON_CONCLUSIONS

MISSING_DATA_SENSITIVITY:
  compare:
    - COMPLETE_CASE
    - MULTIPLE_IMPUTATION
    - WORST_CASE_SCENARIO
    - ALTERNATIVE_MISSINGNESS_MODEL

QUALITY_ASSURANCE:
  preventive:
    - VALIDATED_INSTRUMENTS
    - STANDARD_OPERATING_PROCEDURES
    - TRAINING
    - SCHEMA_CONSTRAINTS
    - AUTOMATED_VALIDATION
    - CALIBRATION

QUALITY_CONTROL:
  detective:
    - RANGE_CHECK
    - DUPLICATE_CHECK
    - OUTLIER_CHECK
    - CONSISTENCY_CHECK
    - SAMPLE_AUDIT

REPRODUCIBILITY:
  preserve:
    - RAW_DATA
    - DATA_VERSION
    - SCHEMA_VERSION
    - VALIDATION_RULES
    - CLEANING_CODE
    - TRANSFORMATION_LOG
    - RANDOM_SEEDS
    - ANALYTIC_DATA_VERSION

APPLICATIONS:
  DATA_QUALITY -> SUPPORTS -> SCIENTIFIC_STATISTICS
  DATA_QUALITY -> SUPPORTS -> MEASUREMENT_ANALYSIS
  DATA_QUALITY -> SUPPORTS -> EXPERIMENTAL_RESEARCH
  DATA_QUALITY -> SUPPORTS -> MACHINE_LEARNING
  DATA_QUALITY -> SUPPORTS -> DATA_INTEGRATION
  DATA_QUALITY -> SUPPORTS -> QUALITY_CONTROL
  DATA_QUALITY -> SUPPORTS -> DECISION_MAKING

MODEL_CONNECTIONS:
  DATA_QUALITY -> CONNECTS_TO -> SCIENTIFIC_STATISTICS
  DATA_QUALITY -> CONNECTS_TO -> MEASUREMENT_UNCERTAINTY
  DATA_QUALITY -> CONNECTS_TO -> ERROR_DISTRIBUTION
  DATA_QUALITY -> CONNECTS_TO -> OUTLIER_DETECTION
  DATA_QUALITY -> CONNECTS_TO -> SENSITIVITY_ANALYSIS
  DATA_QUALITY -> CONNECTS_TO -> MACHINE_LEARNING_STATISTICS

SEMANTIC_EQUIVALENCE:
  "data quality"
    ~= "fitness of data for specified intended use"

  "data completeness"
    ~= "degree to which required data values are present"

  "data validity"
    ~= "conformance of data to specified rules and domains"

  "data lineage"
    ~= "recorded transformation path from source to derived data"

  EXCLUDE_EQUIVALENCE:
    DATA_QUALITY != DATA_ACCURACY_ALONE
    COMPLETENESS != ACCURACY
    VALIDITY != ACCURACY
    OUTLIER != DATA_ERROR
    DUPLICATE != SIMILAR_RECORD
    RAW_DATA != CLEANED_DATA
    LARGE_DATASET != REPRESENTATIVE_DATASET
    DATA_QUALITY != LOW_MEASUREMENT_UNCERTAINTY
    REPRODUCIBILITY != DATA_CORRECTNESS

VALIDATION_RULES:
  RULE_1:
    INTENDED_USE_AND_QUALITY_REQUIREMENTS_MUST_BE_DEFINED
  RULE_2:
    SCHEMA_DATA_TYPES_UNITS_AND_ALLOWED_DOMAINS_MUST_BE_EXPLICIT
  RULE_3:
    MISSING_VALUES_AND_SPECIAL_CODES_MUST_BE_SEMANTICALLY_DISTINGUISHED
  RULE_4:
    OUTLIERS_MUST_NOT_BE_AUTOMATICALLY_CLASSIFIED_AS_ERRORS
  RULE_5:
    CORRECTIONS_AND_TRANSFORMATIONS_MUST_PRESERVE_AUDIT_TRAIL
  RULE_6:
    DUPLICATE_AND_ENTITY_RESOLUTION_RULES_MUST_BE_DOCUMENTED
  RULE_7:
    REPRESENTATIVENESS_AND_DISTRIBUTION_SHIFT_MUST_BE_ASSESSED_WHEN_GENERALISATION_IS_REQUIRED
  RULE_8:
    DATA_PROVENANCE_AND_LINEAGE_MUST_BE_PRESERVED

CONFIDENCE:
  ontology_identity: 0.99
  quality_dimensions_structure: 0.99
  validation_structure: 0.99
  provenance_lineage_structure: 0.99
  statistical_quality_relationship: 0.99
  machine_learning_quality_structure: 0.99
  semantic_distinctions: 0.99

UPDATE_MODEL:
  IF:
    NEW_DATA_ARRIVES
  THEN:
    UPDATE:
      DATA_PROFILE
      COMPLETENESS
      VALIDITY
      UNIQUENESS
      DISTRIBUTION_SUMMARIES
    RERUN:
      VALIDATION_RULES

  IF:
    SCHEMA_OR_VALIDATION_RULES_CHANGE
  THEN:
    UPDATE:
      QUALITY_REQUIREMENTS
    REVALIDATE:
      DATASET
    RECORD:
      RULE_VERSION

  IF:
    DATA_CORRECTION_OCCURS
  THEN:
    PRESERVE:
      ORIGINAL_VALUE
      CORRECTED_VALUE
      REASON
      METHOD
      VERSION
    RECOMPUTE:
      QUALITY_METRICS

  PRESERVE:
    RAW_DATA_VERSION
    DATASET_IDENTIFIER
    SCHEMA_VERSION
    VARIABLE_DEFINITIONS
    UNITS
    VALIDATION_RULES
    QUALITY_THRESHOLDS
    CLEANING_HISTORY
    DATA_LINEAGE
    DATA_PROVENANCE

AI_CONSUMPTION:
  parse_as: FORMAL_SEMANTIC_SPECIFICATION
  execution_mode: RECURSIVE_REASONING

  preferred_reasoning_order:
    - IDENTIFY_INTENDED_DATA_USE
    - IDENTIFY_SCHEMA_VARIABLES_AND_UNITS
    - PROFILE_DATASET
    - ASSESS_COMPLETENESS
    - ASSESS_VALIDITY_AND_CONSISTENCY
    - ASSESS_UNIQUENESS_AND_REFERENTIAL_INTEGRITY
    - ASSESS_ACCURACY_AGAINST_AVAILABLE_REFERENCES
    - REVIEW_OUTLIERS_AND_DATA_ERRORS_SEPARATELY
    - ASSESS_REPRESENTATIVENESS_AND_SHIFT
    - VERIFY_PROVENANCE_AND_LINEAGE
    - APPLY_DOCUMENTED_CORRECTIONS
    - RECOMPUTE_QUALITY_METRICS
    - RETURN_DATA_QUALITY_INTERPRETATION

  constraints:
    - PRESERVE_RAW_DATA
    - PRESERVE_DATA_SEMANTICS
    - PRESERVE_UNITS
    - DISTINGUISH_OUTLIERS_FROM_ERRORS
    - DISTINGUISH_COMPLETENESS_FROM_ACCURACY
    - DOCUMENT_ALL_CORRECTIONS
    - ASSESS_FITNESS_RELATIVE_TO_INTENDED_USE
    - PRESERVE_DATA_PROVENANCE

FLOW_HASH:
  canonical_entity: "data-quality"
  namespace: "probability-statistics/scientific-statistical-methods"
  version: "4.0"
  identity_basis:
    - PAGE_URL
    - CANONICAL_ENTITY
    - ENTITY_TYPE
    - QUALITY_DIMENSIONS
    - VALIDATION_RULES
    - PROVENANCE_AND_LINEAGE
    - FITNESS_FOR_PURPOSE

END_FLOW