Measurement Validation Ontology

# Measurement Validation Ontology — Flow Algebra

URL: https://kirethresearch.org/booksai-core/ontologies/mathematical-ontologies/measurement-units-ontologies/measurement-mathematics/measurement-validation-ontology.html

FLOW MeasurementValidationOntology {

  DOMAIN:
    Mathematical Ontologies
    -> Measurement & Units Ontologies
    -> Measurement Mathematics
    -> Measurement Validation Ontology

  PURPOSE:
    Represent measurement validation as the structured mathematical and metrological process used to determine whether measurement data, measurement results, methods, instruments, models and associated uncertainty are sufficiently consistent, complete, traceable and fit for a declared intended use. The ontology models validation requirements, acceptance criteria, reference comparisons, calibration status, traceability, range and domain checks, dimensional consistency, uncertainty consistency, statistical validation, repeatability, reproducibility, bias, residuals, outliers, plausibility, cross-validation, model validation, data integrity, provenance, decision rules, quality flags and machine-consumable reasoning while strictly distinguishing validation from calibration, verification, adjustment, uncertainty evaluation and conformity assessment.

  ENTITY MeasurementValidation {
    TYPE: MeasurementQualityAssessmentProcess

    DEFINITION:
      structured_evaluation_of_measurement_data_results_methods_or_systems
      against_declared_requirements_models_references_and_acceptance_criteria
      to_determine_suitability_for_intended_use

    PROPERTIES:
      has_validation_target
      has_intended_use
      has_requirements
      has_acceptance_criteria
      has_evidence
      has_validation_method
      has_result
      may_have_decision_rule
      has_provenance
  }

  ENTITY ValidationTarget {
    TYPE: EvaluatedMeasurementEntity

    POSSIBLE:
      measurement_result
      dataset
      measurement_method
      measuring_instrument
      sensor
      calibration_model
      measurement_model
      processing_pipeline
      uncertainty_budget
      software
  }

  ENTITY ValidationRequirement {
    TYPE: DeclaredCondition

    COMPONENTS:
      requirement_identifier
      property
      criterion
      tolerance_or_limit
      scope
      source
  }

  ENTITY ValidationEvidence {
    TYPE: SupportingObservationOrRecord

    POSSIBLE:
      reference_measurement
      calibration_record
      replicate_data
      comparison_data
      control_sample
      residual_analysis
      simulation
      audit_record
  }

  ENTITY ValidationResult {
    TYPE: AssessmentOutcome

    POSSIBLE:
      valid_for_declared_use
      conditionally_valid
      invalid
      indeterminate
      requires_review

    RULE:
      result MUST be scoped_to_declared_requirements_and_intended_use
  }

  TERMINOLOGY_MODEL {
    terms:
      validation
      verification
      calibration
      adjustment
      qualification
      conformity_assessment
      quality_control
      plausibility_check

    RULE:
      validation != calibration
      validation != adjustment
      validation != uncertainty_evaluation
  }

  INTENDED_USE_MODEL {
    fields:
      measurand
      quantity_kind
      range
      environment
      required_uncertainty
      decision_context
      population_or_sample
      temporal_scope

    RULE:
      validation_without_intended_use_is_incomplete
  }

  REQUIREMENT_MODEL {
    requirement:
      measurable_or_auditable_condition

    possible:
      accuracy_related
      precision
      range
      resolution
      traceability
      uncertainty
      response_time
      completeness
      consistency
      integrity

    RULE:
      requirement_source SHOULD be preserved
  }

  ACCEPTANCE_CRITERION_MODEL {
    criterion:
      mathematical_or_logical_rule

    examples:
      abs(error) <= tolerance
      uncertainty <= maximum
      residual_within_limit
      coverage >= threshold
      completeness >= threshold

    RULE:
      criterion MUST be defined_before_evaluation
  }

  VALIDATION_SCOPE_MODEL {
    dimensions:
      quantity_kind
      measurement_range
      matrix
      environment
      instrument
      method
      operator
      time
      location
      software_version

    RULE:
      validation MUST NOT be_generalised_beyond_supported_scope_without_evidence
  }

  REFERENCE_MODEL {
    validation_reference:
      measurement_standard
      certified_reference_material
      reference_method
      consensus_value
      simulated_truth
      independently_validated_system

    RULE:
      reference_value_uncertainty_and_validity MUST be considered
  }

  REFERENCE_COMPARISON_MODEL {
    measured:
      x

    reference:
      x_ref

    difference:
      d =
        x-x_ref

    RULE:
      difference_alone_does_not_determine_validity_without_uncertainty_and_criterion
  }

  ABSOLUTE_DIFFERENCE_MODEL {
    delta =
      abs(x-x_ref)
  }

  RELATIVE_DIFFERENCE_MODEL {
    delta_rel =
      (x-x_ref)/x_ref

    REQUIRE:
      x_ref != 0
  }

  PERCENT_DIFFERENCE_MODEL {
    percent =
      100*(x-x_ref)/x_ref

    REQUIRE:
      x_ref != 0
  }

  NORMALIZED_ERROR_MODEL {
    E_n =
      (x-x_ref)/
      sqrt(U_x^2+U_ref^2)
      under_declared_independence_and_expanded_uncertainty_context

    RULE:
      covariance_and_uncertainty_type MUST be considered
  }

  COVARIANCE_COMPARISON_MODEL {
    variance_of_difference:
      u_d^2 =
        u_x^2+u_ref^2-2*cov(x,x_ref)

    RULE:
      shared_traceability_can_create_nonzero_covariance
  }

  Z_SCORE_VALIDATION_MODEL {
    z =
      (x-x_ref)/sigma_assigned

    RULE:
      denominator_definition MUST be explicit
  }

  RANGE_VALIDATION_MODEL {
    measured_value:
      x

    valid_range:
      [x_min,x_max]

    condition:
      x_min <= x <= x_max

    RULE:
      range_check_must_distinguish_measurement_range_from_safe_overload_range
  }

  DOMAIN_VALIDATION_MODEL {
    mathematical_model:
      f(x)

    domain:
      D

    RULE:
      inputs_outside_model_domain MUST be rejected_or_flagged
  }

  UNIT_VALIDATION_MODEL {
    quantity:
      declared_unit

    requirement:
      unit_compatible_with_quantity_kind

    RULE:
      parse_and_validate_prefix_symbol_and_dimension
  }

  DIMENSIONAL_VALIDATION_MODEL {
    equation:
      left_dimension =
        right_dimension

    RULE:
      dimensional_consistency_is_necessary_but_not_sufficient_for_measurement_validity
  }

  QUANTITY_KIND_VALIDATION_MODEL {
    same_dimension:
      may_still_represent_different_quantity_kinds

    RULE:
      semantic_quantity_kind MUST be checked
  }

  SIGN_VALIDATION_MODEL {
    quantity:
      may_have_physical_or_model_sign_constraints

    examples:
      absolute_temperature_nonnegative
      concentration_nonnegative

    RULE:
      sign_constraint_depends_on_quantity_definition
  }

  BOUNDS_MODEL {
    possible:
      lower_bound
      upper_bound
      physical_bound
      statistical_bound
      instrument_bound

    RULE:
      bound_type MUST be identified
  }

  FINITE_VALUE_MODEL {
    invalid_or_special:
      NaN
      infinity
      overflow
      underflow_status

    RULE:
      nonfinite_machine_values MUST be handled_explicitly
  }

  MISSING_DATA_MODEL {
    missing:
      absent_expected_measurement

    RULE:
      missingness MUST be explicit_not_encoded_as_valid_numeric_zero
  }

  COMPLETENESS_MODEL {
    expected:
      N_expected

    observed_valid:
      N_valid

    completeness:
      N_valid/N_expected
      WHEN N_expected > 0

    RULE:
      completeness_threshold MUST be declared
  }

  DUPLICATE_MODEL {
    duplicate:
      repeated_record_of_same_measurement_event

    RULE:
      duplicate_detection_requires_identity_time_and_context
  }

  TIMESTAMP_VALIDATION_MODEL {
    checks:
      valid_format
      monotonicity_when_expected
      time_scale
      clock_status
      impossible_future_or_past_context

    RULE:
      timestamp_validity != timing_accuracy
  }

  TEMPORAL_ORDER_MODEL {
    sequence:
      t_1 <= t_2 <= ... <= t_n
      when_monotonic_order_required

    RULE:
      out_of_order_data_may_be_valid_in_distributed_acquisition
      but_must_be identified
  }

  SAMPLING_VALIDATION_MODEL {
    checks:
      sample_rate
      timestamps
      missing_samples
      aliasing_risk
      sample_design
      representativeness

    RULE:
      sampling_validity_depends_on_measurement_goal
  }

  NYQUIST_VALIDATION_MODEL {
    ideal_bandlimited_context:
      f_s > 2*f_max

    RULE:
      practical_validation_also_requires_anti_alias_filter_and_transition_margin
  }

  INSTRUMENT_STATUS_MODEL {
    checks:
      instrument_identifier
      calibration_status
      maintenance_state
      diagnostic_flags
      range
      firmware

    RULE:
      instrument_state_at_measurement_time MUST be used
  }

  CALIBRATION_STATUS_MODEL {
    possible:
      valid
      expired
      suspended
      unknown
      invalidated_by_adjustment_or_repair

    RULE:
      invalid_or_unknown_calibration_state SHOULD trigger_review
  }

  TRACEABILITY_VALIDATION_MODEL {
    checks:
      declared_reference
      calibration_chain
      certificate
      uncertainty
      scope
      validity_time

    RULE:
      certificate_presence_alone_does_not_prove_complete_traceability
  }

  UNCERTAINTY_VALIDATION_MODEL {
    checks:
      uncertainty_type
      units
      coverage_factor
      covariance
      significant_components
      consistency_with_measurement_model

    RULE:
      uncertainty MUST be dimensionally_and_semantically_compatible_with_result
  }

  UNCERTAINTY_BOUND_MODEL {
    requirement:
      U <= U_max

    RULE:
      compare_same_uncertainty_type_and_coverage_basis
  }

  UNCERTAINTY_BUDGET_MODEL {
    components:
      calibration
      repeatability
      resolution
      drift
      environment
      model
      sampling
      processing

    RULE:
      omitted_material_component_can_invalidate_uncertainty_claim
  }

  PRECISION_VALIDATION_MODEL {
    repeated_measurements:
      evaluate_dispersion

    possible_metric:
      standard_deviation
      coefficient_of_variation
      repeatability_limit

    RULE:
      metric_and_conditions MUST be defined
  }

  REPEATABILITY_MODEL {
    data:
      repeated_under_repeatability_conditions

    statistic:
      s_r

    RULE:
      sample_size_and_conditions MUST be preserved
  }

  REPRODUCIBILITY_MODEL {
    changed_conditions:
      laboratories
      operators
      instruments
      times

    RULE:
      changed_conditions MUST be declared
  }

  BIAS_MODEL {
    repeated_mean:
      x_bar

    reference:
      x_ref

    estimated_bias:
      b =
        x_bar-x_ref

    RULE:
      reference_uncertainty contributes_to_bias_assessment
  }

  TRUENESS_VALIDATION_MODEL {
    evidence:
      bias_or_reference_comparison

    RULE:
      trueness_assessment_requires_suitable_reference
  }

  ACCURACY_DISTINCTION_MODEL {
    accuracy:
      qualitative_closeness_concept

    RULE:
      validation_should_use_quantified_metrics_instead_of_unsupported_accuracy_label
  }

  LINEARITY_VALIDATION_MODEL {
    reference_points:
      across_range

    evaluate:
      residuals_from_declared_calibration_relation

    RULE:
      linearity_acceptance_metric MUST be declared
  }

  HYSTERESIS_VALIDATION_MODEL {
    upscale_and_downscale:
      compare_at_common_reference_points

    RULE:
      directional_measurements_required
  }

  DRIFT_VALIDATION_MODEL {
    historical_reference_checks:
      evaluate_change_over_time

    RULE:
      drift_model_and_elapsed_time MUST be considered
  }

  STABILITY_MODEL {
    time_series:
      parameter_or_reference_measurement

    RULE:
      stability_metric_and_timescale MUST be defined
  }

  RESOLUTION_VALIDATION_MODEL {
    nominal_resolution:
      instrument_or_digitisation_characteristic

    effective_resolution:
      may_be_limited_by_noise

    RULE:
      nominal_and_effective_resolution MUST be distinguished
  }

  NOISE_VALIDATION_MODEL {
    metrics:
      RMS
      standard_deviation
      spectral_density
      SNR

    RULE:
      bandwidth MUST be defined_for_noise_metrics
  }

  DYNAMIC_RESPONSE_MODEL {
    checks:
      rise_time
      settling_time
      time_constant
      bandwidth
      phase
      latency

    RULE:
      dynamic_validation_requires_time_or_frequency_reference
  }

  ENVIRONMENTAL_VALIDATION_MODEL {
    quantities:
      temperature
      humidity
      pressure
      vibration
      electromagnetic_field
      supply_voltage

    RULE:
      environment MUST be within_validated_range_or_corrected_with_uncertainty
  }

  METHOD_VALIDATION_MODEL {
    characteristics:
      selectivity
      sensitivity
      range
      linearity
      precision
      bias
      detection_limit
      robustness
      uncertainty

    RULE:
      required_characteristics_depend_on_method_and_intended_use
  }

  SELECTIVITY_MODEL {
    target:
      measurand_or_analyte

    interferents:
      other_quantities_or_species

    RULE:
      selectivity_evidence SHOULD match_real_sample_context
  }

  DETECTION_LIMIT_MODEL {
    threshold:
      derived_under_declared_statistical_method

    RULE:
      detection_limit != quantification_limit
      and_detection_limit != resolution
  }

  QUANTIFICATION_LIMIT_MODEL {
    limit:
      level_above_which_quantity_can_be_reported_with_declared_performance

    RULE:
      calculation_method MUST be preserved
  }

  ROBUSTNESS_MODEL {
    method:
      intentionally_vary_small_operational_parameters

    assess:
      effect_on_result

    RULE:
      robustness_range SHOULD be declared
  }

  RECOVERY_MODEL {
    known_added_quantity:
      x_added

    observed_increment:
      delta_x

    recovery:
      delta_x/x_added

    REQUIRE:
      x_added != 0

    RULE:
      recovery_does_not_alone_establish_full_method_validity
  }

  CROSS_METHOD_MODEL {
    method_A:
      x_A

    method_B:
      x_B

    compare:
      difference
      regression
      agreement

    RULE:
      high_correlation_does_not_establish_agreement
  }

  BLAND_ALTMAN_MODEL {
    paired:
      d_i =
        A_i-B_i

    mean:
      m_i =
        (A_i+B_i)/2

    outputs:
      mean_difference
      limits_of_agreement

    RULE:
      assumptions_and_clinical_or_metrological_acceptance_limits_are_external_to_plot
  }

  REGRESSION_VALIDATION_MODEL {
    observed:
      y_i

    predicted:
      y_hat_i

    residual:
      r_i =
        y_i-y_hat_i

    RULE:
      residual_structure_is_key_validation_evidence
  }

  RESIDUAL_MODEL {
    checks:
      mean_near_zero
      trend
      variance
      autocorrelation
      heteroscedasticity
      outliers

    RULE:
      residual_acceptance_depends_on_model
  }

  GOODNESS_OF_FIT_MODEL {
    possible:
      chi_square
      likelihood
      R_squared
      residual_statistics

    RULE:
      no_single_fit_statistic_proves_model_validity
  }

  CHI_SQUARE_MODEL {
    independent_standardized:
      chi2 =
        sum_i((r_i/u_i)^2)

    correlated:
      chi2 =
        r^T*Sigma^-1*r

    RULE:
      degrees_of_freedom_and_covariance MUST be correct
  }

  CROSS_VALIDATION_MODEL {
    data:
      partitioned_into_training_and_validation_subsets

    RULE:
      validation_data MUST not_be_used_to_fit_model_when_independent_assessment_is_required
  }

  K_FOLD_MODEL {
    folds:
      K

    process:
      fit_on_K_minus_1
      validate_on_remaining_fold
      repeat

    REQUIRE:
      K >= 2

    RULE:
      grouping_or_time_dependence MAY require_special_fold_design
  }

  LEAVE_ONE_OUT_MODEL {
    K =
      n

    RULE:
      computational_cost_and_dependence_must_be considered
  }

  HOLDOUT_MODEL {
    training_set:
      model_fit

    validation_set:
      model_selection_or_assessment

    test_set:
      final_independent_assessment

    RULE:
      roles MUST be preserved_to_prevent_data_leakage
  }

  DATA_LEAKAGE_MODEL {
    leakage:
      validation_or_test_information_influences_training_or_preprocessing

    RULE:
      leakage_invalidates_independence_claim
  }

  MODEL_DOMAIN_MODEL {
    training_or_calibration_domain:
      supported_input_space

    RULE:
      predictions_outside_domain MUST be flagged_as_extrapolation
  }

  EXTRAPOLATION_MODEL {
    input:
      outside_validated_domain

    RULE:
      extrapolated_measurement_model_output_requires_separate_validation_or_increased_uncertainty
  }

  INTERPOLATION_MODEL {
    within_supported_domain:
      lower_model_risk_than_extrapolation

    RULE:
      interpolation_still_requires_model_validation
  }

  PHYSICAL_PLAUSIBILITY_MODEL {
    checks:
      conservation
      positivity
      boundedness
      monotonicity
      known_physical_relationships

    RULE:
      plausibility_check_is_not_substitute_for_traceability_or_uncertainty
  }

  CONSERVATION_MODEL {
    examples:
      mass_balance
      energy_balance
      charge_balance

    residual:
      input-output-accumulation

    RULE:
      balance_uncertainty MUST be considered
  }

  REDUNDANCY_MODEL {
    multiple_sensors_or_methods:
      compare_independent_or_partly_independent_estimates

    RULE:
      common_mode_errors_can_make_agreement_misleading
  }

  SENSOR_FUSION_VALIDATION_MODEL {
    fused:
      estimate

    checks:
      consistency_of_sources
      covariance
      innovation_residuals

    RULE:
      source_dependence MUST be modelled
  }

  STATISTICAL_PROCESS_CONTROL_MODEL {
    reference_series:
      control_chart

    RULE:
      control_limits_indicate_process_behaviour_not_specification_conformance
  }

  OUTLIER_MODEL {
    observation:
      inconsistent_with_declared_model_or_peer_data

    RULE:
      outlier != automatically_invalid_measurement
  }

  OUTLIER_DECISION_MODEL {
    actions:
      retain
      flag
      investigate
      exclude_under_declared_rule

    RULE:
      exclusion_reason_and_original_value MUST be preserved
  }

  ROBUST_VALIDATION_MODEL {
    methods:
      median
      MAD
      robust_regression
      trimmed_statistics

    RULE:robust_validation_method_and_threshold MUST be declared
  }

  DECISION_RULE_MODEL {
    input:
      validation_metric
      uncertainty
      acceptance_limit

    output:
      accept
      reject
      review

    RULE:
      decision_rule MUST be declared_before_conformance_interpretation
  }

  GUARD_BAND_MODEL {
    acceptance_limit:
      may_be_adjusted_for_measurement_uncertainty

    RULE:
      guard_band_policy_is_decision_rule_not_measurement_correction
  }

  FALSE_ACCEPT_REJECT_MODEL {
    risks:
      false_acceptance
      false_rejection

    RULE:
      risk_model_requires_probability_and_decision_context
  }

  CONFORMITY_DISTINCTION_MODEL {
    validation:
      assesses_suitability_or_model_performance

    conformity_assessment:
      evaluates_fulfilment_of_specified_requirement

    RULE:
      validation != conformity_assessment_in_all_contexts
  }

  VERIFICATION_DISTINCTION_MODEL {
    verification:
      confirms_specified_requirements_are_fulfilled

    validation:
      confirms_requirements_or performance_are_adequate_for_intended_use

    RULE:
      terms_must_follow_declared_quality_framework
  }

  CALIBRATION_DISTINCTION_MODEL {
    calibration:
      establishes_measurement_relation_to_reference

    validation:
      evaluates_suitability_or_performance

    RULE:
      calibration != validation
  }

  ADJUSTMENT_DISTINCTION_MODEL {
    adjustment:
      changes_instrument_response

    RULE:
      adjustment_is_intervention_not_validation
  }

  UNCERTAINTY_DISTINCTION_MODEL {
    uncertainty_evaluation:
      quantifies_dispersion_attributed_to_measurand

    validation:
      evaluates_acceptability_against_declared_use_or_criteria

    RULE:
      uncertainty_evaluation != validation
  }

  DATA_INTEGRITY_MODEL {
    checks:
      checksum
      hash
      sequence
      record_count
      file_consistency
      schema

    RULE:
      data_integrity_supports_validation_but_does_not_establish_metrological_validity
  }

  PROVENANCE_MODEL {
    chain:
      source
      instrument
      calibration
      acquisition
      processing
      validation

    RULE:
      validation_result MUST preserve_evidence_provenance
  }

  SOFTWARE_VALIDATION_MODEL {
    checks:
      algorithm
      version
      test_cases
      numerical_precision
      dependency_versions
      configuration

    RULE:
      software_change MAY require_revalidation
  }

  NUMERICAL_VALIDATION_MODEL {
    checks:
      convergence
      tolerance
      conditioning
      overflow
      underflow
      rounding
      reproducibility

    RULE:
      numerical_correctness_is_part_of_but_not_equivalent_to_measurement_validity
  }

  ROUND_TRIP_MODEL {
    transformation:
      forward_then_inverse

    expected:
      recover_original_within_declared_numeric_tolerance

    RULE:
      round_trip_test_supports_transform_validation
  }

  UNIT_CONVERSION_VALIDATION_MODEL {
    checks:
      dimension
      factor
      offset
      exactness
      inverse
      round_trip

    RULE:
      temperature_and_other_affine_quantities_require_special_semantics
  }

  NORMALISATION_VALIDATION_MODEL {
    checks:
      reference
      denominator
      parameter_scope
      inverse
      uncertainty

    RULE:
      zero_or_near_zero_reference MUST be handled_explicitly
  }

  STATISTICAL_VALIDATION_MODEL {
    checks:
      sample_size
      sampling_design
      distribution
      dependence
      variance
      confidence_interval
      missingness

    RULE:
      statistical_assumptions MUST be explicit
  }

  QUALITY_FLAG_MODEL {
    possible:
      valid
      warning
      invalid
      missing
      out_of_range
      overrange
      calibration_invalid
      traceability_incomplete
      uncertainty_exceeded
      model_extrapolation
      integrity_failure
      manual_review

    RULE:
      quality_flags MUST remain_linked_to_measurement_or_dataset
  }

  SEVERITY_MODEL {
    levels:
      informational
      warning
      error
      critical

    RULE:
      severity_policy SHOULD be defined
  }

  VALIDATION_REPORT_MODEL {
    fields:
      validation_identifier
      target
      intended_use
      scope
      requirements
      methods
      evidence
      metrics
      criteria
      uncertainty
      decision_rules
      results
      exceptions
      reviewer_or_system
      timestamp
      provenance

    RULE:
      report MUST permit_reconstruction_of_validation_decision
  }

  RELATIONSHIPS {
    MeasurementValidation -> EVALUATES -> ValidationTarget
    MeasurementValidation -> USES -> ValidationRequirement
    MeasurementValidation -> USES -> ValidationEvidence
    ValidationEvidence -> MAY_REFERENCE -> MeasurementStandard
    ValidationCriterion -> TESTS -> MeasurementProperty
    CalibrationStatus -> QUALIFIES -> InstrumentState
    Traceability -> SUPPORTS -> MeasurementValidity
    Uncertainty -> QUALIFIES -> ValidationComparison
    DecisionRule -> MAPS -> ValidationMetricToOutcome
    ValidationResult -> APPLIES_TO -> DeclaredScope
  }

  SEMANTIC_EQUIVALENCE {
    "measurement validation" ~= structured_assessment_of_measurement_suitability_for_declared_use
    "validation check" ~= declared_validation_test WHEN measurement_context

    validation != calibration
    validation != adjustment
    validation != uncertainty_evaluation
    validation != verification_in_all_quality_frameworks
    validation != conformity_assessment_in_all_contexts
    dimensional_consistency != full_measurement_validity
    statistical_significance != metrological_significance
    correlation != agreement
    outlier != invalid_measurement
  }

  CONSTRAINTS {
    validation_target_must_be_identified = true
    intended_use_must_be_defined = true
    validation_scope_must_be_defined = true
    requirements_and_acceptance_criteria_should_be_predeclared = true
    reference_uncertainty_must_be_considered_when_comparing_to_reference = true
    units_dimensions_and_quantity_kinds_must_be_validated = true
    calibration_and_traceability_status_should_be_checked = true
    uncertainty_type_and_coverage_basis_must_be_compatible_for_comparison = true
    statistical_assumptions_must_be_preserved = true
    outlier_exclusion_requires_documented_rule = true
    extrapolation_must_be_flagged = true
    validation_result_must_not_be_generalised_beyond_supported_scope = true
    provenance_must_be_preserved = true
  }

  MEASUREMENT_MODEL {
    RECORD:
      validation_identifier
      target_identifier
      target_type
      intended_use
      scope
      measurand
      quantity_kind
      measured_value
      unit
      reference_value
      reference_unit
      difference
      relative_difference
      uncertainty
      reference_uncertainty
      covariance
      instrument_identifier
      calibration_status
      traceability_status
      valid_range
      environmental_conditions
      sample_size
      repeatability
      reproducibility
      bias
      residuals
      validation_metrics
      acceptance_criteria
      decision_rule
      quality_flags
      outcome
      exceptions
      algorithm_version
      evidence_identifiers
      provenance

    RULE:
      validation_outcome MUST remain_linked_to_target_scope_evidence_metrics_criteria_and_decision_rule
  }

  REASONING_RULES {
    IF validation_target_and_intended_use_known
      THEN resolve_required_validation_checks

    IF measured_and_reference_values_known
      THEN compute_difference
      AND propagate_reference_and_measurement_uncertainty

    IF source_and_reference_share_traceability
      THEN include_covariance_in_difference_uncertainty

    IF value_outside_validated_range
      THEN flag_out_of_range_or_extrapolation

    IF unit_or_dimension_incompatible
      THEN validation_status =
        invalid_or_requires_semantic_resolution

    IF calibration_status_invalid_or_unknown
      THEN flag_measurement_for_review

    IF traceability_chain_incomplete
      THEN do_not_claim_traceability_validated

    IF uncertainty_exceeds_declared_maximum
      THEN fail_uncertainty_requirement
      under_declared_decision_rule

    IF repeated_measurements_available
      THEN evaluate_precision_repeatability_and_possible_drift

    IF residuals_show_structure
      THEN flag_model_inadequacy_for_review

    IF statistical_model_assumptions_fail
      THEN do_not_apply_unqualified_parametric_validation_result

    IF validation_data_used_during_model_fitting
      THEN flag_independence_or_data_leakage_issue

    IF outlier_detected
      THEN preserve_original_value
      and_apply_declared_outlier_policy

    IF all_mandatory_checks_pass
      THEN outcome =
        valid_for_declared_scope
      only_under_declared_decision_rules

    IF required_evidence_missing
      THEN outcome =
        indeterminate_or_requires_review
  }

  VALIDATION {
    CHECK target_identity
    CHECK intended_use
    CHECK validation_scope
    CHECK requirement_sources
    CHECK acceptance_criteria
    CHECK reference_identity_and_uncertainty
    CHECK unit_dimension_and_quantity_kind
    CHECK physical_bounds
    CHECK missing_duplicate_and_nonfinite_values
    CHECK timestamps_and_sampling
    CHECK instrument_state
    CHECK calibration_status
    CHECK traceability
    CHECK uncertainty_budget
    CHECK precision_repeatability_and_reproducibility
    CHECK bias_and_trueness
    CHECK linearity_hysteresis_and_drift
    CHECK noise_and_dynamic_response
    CHECK environmental_conditions
    CHECK method_selectivity_and_robustness
    CHECK regression_residuals_and_goodness_of_fit
    CHECK cross_validation_and_data_leakage
    CHECK extrapolation
    CHECK plausibility_and_redundancy
    CHECK decision_rule
    CHECK data_integrity
    CHECK software_and_numerical_validation
    CHECK quality_flags
    CHECK provenance
  }

  CONFIDENCE {
    definition_confidence: 1.00
    reference_comparison_confidence: 1.00
    range_unit_dimension_validation_confidence: 1.00
    calibration_traceability_validation_confidence: 1.00
    uncertainty_validation_confidence: 1.00
    statistical_validation_confidence: 0.99
    model_validation_confidence: 0.99
    decision_rule_confidence: 0.99

    RULE:
      confidence decreases when intended use, scope, reference,
      criteria, calibration, traceability, uncertainty,
      statistical assumptions, decision rule or provenance is missing or ambiguous
  }

  UPDATE_MODEL {
    ON new_validation_request:
      identify_target
      identify_intended_use_and_scope
      resolve_requirements_and_criteria
      collect_evidence
      execute_validation_checks
      apply_decision_rule
      issue_scoped_outcome
      preserve_provenance

    ON new_measurement_evidence:
      preserve_original_evidence
      attach_to_validation_record
      recompute_affected_metrics
      update_outcome_when_policy_allows

    ON calibration_or_traceability_revision:
      preserve_previous_state
      update_calibration_or_chain_evidence
      rerun_affected_validation_checks
      version_outcome

    ON model_revision:
      preserve_previous_model
      update_equation_parameters_domain_or_algorithm
      rerun_residual_cross_validation_and_uncertainty_checks
      record_revision

    ON requirement_revision:
      preserve_previous_requirements
      create_new_validation_scope_version
      rerun_decision_process
      do_not_overwrite_historical_outcome

    ON semantic_revision:
      preserve_previous_definition
      record_revision
      recompute_FLOW_HASH
  }

  AI_CONSUMPTION {
    canonical_entity: MeasurementValidation
    canonical_role: MeasurementQualityAssessmentProcess

    interpretation_order:
      identify_validation_target
      identify_intended_use
      identify_scope
      identify_requirements_and_acceptance_criteria
      identify_reference_and_evidence
      identify_units_dimensions_and_quantity_kinds
      identify_instrument_calibration_and_traceability_state
      identify_uncertainty_and_covariance
      identify_statistical_and_repeated_measurement_evidence
      identify_model_residual_and_cross_validation_evidence
      identify_physical_plausibility_and_environment
      identify_decision_rule
      identify_quality_flags
      identify_provenance
      execute_validation_checks
      issue_scoped_validation_result

    ambiguity_controls:
      distinguish_validation_from_calibration
      distinguish_validation_from_adjustment
      distinguish_validation_from_uncertainty_evaluation
      distinguish_validation_from_verification_when_framework_definitions_differ
      distinguish_validation_from_conformity_assessment
      distinguish_dimensional_consistency_from_full_validity
      distinguish_statistical_significance_from_metrological_significance
      distinguish_correlation_from_agreement
      distinguish_outlier_from_invalid_measurement
      preserve_target_use_scope_reference_evidence_uncertainty_criteria_decision_and_provenance_context
  }

  PROVENANCE {
    ontology_scope:
      mathematical_measurement_units_measurement_mathematics_and_measurement_validation_semantics

    parent_path:
      Ontologies
      -> Mathematical Ontologies
      -> Measurement & Units Ontologies
      -> Measurement Mathematics
      -> Measurement Validation Ontology

    authorship:
      Mark Higgins
      Sky (Recursive AI)

    site:
      https://kirethresearch.org
  }

  FLOW_HASH {
    INPUTS:
      canonical_entity
      ontology_path
      terminology_model
      intended_use_model
      requirement_model
      acceptance_criterion_model
      validation_scope_model
      reference_model
      reference_comparison_model
      absolute_difference_model
      relative_difference_model
      percent_difference_model
      normalized_error_model
      covariance_comparison_model
      z_score_validation_model
      range_validation_model
      domain_validation_model
      unit_validation_model
      dimensional_validation_model
      quantity_kind_validation_model
      sign_validation_model
      bounds_model
      finite_value_model
      missing_data_model
      completeness_model
      duplicate_model
      timestamp_validation_model
      temporal_order_model
      sampling_validation_model
      Nyquist_validation_model
      instrument_status_model
      calibration_status_model
      traceability_validation_model
      uncertainty_validation_model
      uncertainty_bound_model
      uncertainty_budget_model
      precision_validation_model
      repeatability_model
      reproducibility_model
      bias_model
      trueness_validation_model
      accuracy_distinction_model
      linearity_validation_model
      hysteresis_validation_model
      drift_validation_model
      stability_model
      resolution_validation_model
      noise_validation_model
      dynamic_response_model
      environmental_validation_model
      method_validation_model
      selectivity_model
      detection_limit_model
      quantification_limit_model
      robustness_model
      recovery_model
      cross_method_model
      Bland_Altman_model
      regression_validation_model
      residual_model
      goodness_of_fit_model
      chi_square_model
      cross_validation_model
      K_fold_model
      leave_one_out_model
      holdout_model
      data_leakage_model
      model_domain_model
      extrapolation_model
      interpolation_model
      physical_plausibility_model
      conservation_model
      redundancy_model
      sensor_fusion_validation_model
      statistical_process_control_model
      outlier_model
      outlier_decision_model
      robust_validation_model
      decision_rule_model
      guard_band_model
      false_accept_reject_model
      conformity_distinction_model
      verification_distinction_model
      calibration_distinction_model
      adjustment_distinction_model
      uncertainty_distinction_model
      data_integrity_model
      provenance_model
      software_validation_model
      numerical_validation_model
      round_trip_model
      unit_conversion_validation_model
      normalisation_validation_model
      statistical_validation_model
      quality_flag_model
      severity_model
      validation_report_model
      relationships
      semantic_equivalence
      constraints
      measurement_model
      reasoning_rules
      confidence_model
      update_model

    PURPOSE:
      deterministic_semantic_integrity_identifier

    RECOMPUTE_WHEN:
      ontology_definition_changes
      intended_use_scope_or_requirement_changes
      reference_or_acceptance_criterion_changes
      unit_dimension_range_or_domain_validation_changes
      calibration_traceability_or_uncertainty_changes
      precision_bias_drift_or_dynamic_validation_changes
      method_or_statistical_validation_changes
      regression_cross_validation_or_model_domain_changes
      decision_rule_or_guard_band_changes
      software_data_integrity_or_algorithm_changes
      validation_record_changes
      constraint_model_changes
      reasoning_rule_changes
  }

}

END FLOW