Title: MLIPs Ontology: An Ontology for Machine Learning Interatomic Potentials
An OWL ontology capturing the concepts needed to describe machine learning interatomic potentials: algorithm families, hyperparameters, training datasets with DFT provenance, and published benchmarks.
Version: 0.1.0
Namespace: https://w3id.org/mlips#
Prefix: mlips:
Persistent IRI: https://w3id.org/mlips
Licence: CC BY 4.0
Editors:
Contributors:
A hyperparameter of an MLIP method, with its name, data type, range, and default value.
Subclass of: mlschema:HyperParameter
Hyperparameter
A hyperparameter of an MLIP method, with its name,
data type, range, and default value.
]]>A hyperparameter that has a direct physical meaning (e.g., the cutoff radius bounds the physical interaction range, the minimum distance bounds short-range repulsion). Not disjoint with ArchitecturalHyperparameter: many MLIP hyperparameters are simultaneously architectural and physical.
Subclass of: mlips:Hyperparameter
Physical Hyperparameter
A hyperparameter that has a direct physical
meaning. Not disjoint with ArchitecturalHyperparameter.
A hyperparameter that specifies the architecture of the model (number of layers, number of basis functions, maximum moment level, etc.). Not disjoint with PhysicalHyperparameter.
Subclass of: mlips:Hyperparameter
Architectural Hyperparameter
A hyperparameter that specifies the model
architecture. Not disjoint with PhysicalHyperparameter.
A hyperparameter of the training procedure (optimizer type, learning rate, batch size, loss-term weights, regularization strength). These are parameters of how the model is fitted, not of the model itself.
Subclass of: mlips:Hyperparameter
Training Hyperparameter
A hyperparameter of the training procedure.
A concrete value assigned to a hyperparameter in a specific training run or configuration.
Hyperparameter Setting
A concrete value assigned to a hyperparameter in a
specific training run or configuration.
]]>A software implementation of an MLIP method in a specific library and version.
Subclass of: mlschema:Implementation
Implementation
A software implementation of an MLIP method in a
specific library and version.
]]>A software library or tool used in MLIP research (e.g., ASE, LAMMPS, VASP, GPAW).
Library
A software library or tool used in MLIP research
(e.g., ASE, LAMMPS, VASP, GPAW).
]]>A machine learning interatomic potential method family (e.g., MACE, ACE, HDNNP, NequIP, M3GNet, MTP). An MLIP method specifies a functional form mapping an atomic configuration to an energy, a loss function defined over that form, and a training algorithm used to fit its parameters. MLIPMethod is disjoint from mls:Algorithm: an MLIP method is a named specification of all three components, whereas mls:Algorithm refers to a training/learning procedure (e.g., Adam, L-BFGS).
Disjoint with: mlschema:Algorithm
MLIP Method
A machine learning interatomic potential method family
(e.g., MACE, ACE, HDNNP, NequIP, M3GNet, MTP). Specifies a functional form,
a loss function, and a training algorithm. Disjoint from mls:Algorithm.
]]>A parametrised mapping from an atomic configuration to an energy (and derived quantities such as forces and stresses), specified independently of the values of its parameters. Examples: the MTP moment-tensor polynomial $E(\text{cfg}; \boldsymbol{\xi})$, the MACE message-passing ansatz, the HDNNP sum of atomic neural networks. The functional form is what is being fitted; the trained model records the fitted parameter values.
Functional Form
A parametrised mapping from an atomic configuration to an energy, specified independently of parameter values.
]]>A scalar objective whose minimisation fits the parameters of the functional form. For MLIPs it is typically a weighted sum of mean-squared errors on energies, forces, and stresses, optionally with regularisation terms.
Loss Function
A scalar objective whose minimisation fits the parameters of the functional form.
]]>A type of atomistic simulation: molecular dynamics, Monte Carlo, geometry optimization, phonon calculations, thermodynamic integration, etc.
Simulation Type
A type of atomistic simulation: molecular dynamics,
Monte Carlo, geometry optimization, phonon calculations,
thermodynamic integration, etc.
]]>The concrete artifact produced by fitting an MLIP method on a specific dataset with specific hyperparameter settings---the learned parameters (weights) that can be used for prediction.
Subclass of: mlschema:Model
Trained Model
The concrete artifact produced by fitting an MLIP method
on a specific dataset with specific hyperparameter settings---the learned
parameters (weights) that can be used for prediction.
]]>A concrete fitting activity for an MLIP method: it applies an MLIP method to a training dataset and produces a trained model. MLIPRun is a subclass of prov:Activity but not of mls:Run, because the mls:Run.executes property ranges over mls:Algorithm and mlips:MLIPMethod is disjoint from mls:Algorithm. An optional hasTrainingRun property links an MLIPRun to an underlying mls:Run when the training algorithm is itself modelled in ML-Schema.
Subclass of: prov:Activity
MLIP Run
A concrete fitting activity for an MLIP method: it applies an MLIP method to a training dataset and produces a trained model.
1
1
1
]]>A mathematical representation of the local atomic environment (neighbor positions and types) used by an MLIP algorithm. Different algorithms use different descriptors: moment tensors (MTP), SOAP (GAP), symmetry functions (HDNNP), equivariant message passing (MACE, NequIP), or atomic cluster expansion (ACE).
Atomic Environment Descriptor
A mathematical representation of the local atomic environment used by an MLIP method.
]]>Links a benchmark result to the trained model being evaluated.
evaluates model
Links a benchmark result to the trained model being evaluated.
Links an MLIP run to the MLIP method it fits.
applies method
Links an MLIP run to the MLIP method it fits.
Links an MLIP method to its parametrised functional form.
has functional form
Links an MLIP method to its parametrised functional form.
Links an MLIP method to the loss function whose minimisation fits its parameters.
has loss function
Links an MLIP method to the loss function whose minimisation fits its parameters.
Links an MLIP method to the ML-Schema algorithm used to fit its functional form by minimising its loss function. This is the alignment point with ML-Schema: while MLIPMethod is disjoint from mls:Algorithm, every MLIP method names the training algorithm (e.g., L-BFGS, Adam, stochastic gradient descent) as an mls:Algorithm instance.
mlschema:Algorithm
has training algorithm
Links an MLIP method to the ML-Schema algorithm used to fit its functional form.
Optional link from an MLIP run to the underlying ML-Schema run that records the execution of the training algorithm. Use when the training algorithm is itself modelled in ML-Schema; omit otherwise.
mlschema:Run
has training run
Optional link from an MLIP run to an ML-Schema run recording the training algorithm execution.
Links a setting to the hyperparameter it configures.
for hyperparameter
Links a setting to the hyperparameter it configures.
Links an MLIP method to a hyperparameter it accepts.
has hyperparameter
Links an MLIP method to a hyperparameter it accepts.
Links an MLIP method to a software implementation.
has implementation
Links an MLIP method to a software implementation.
Links an implementation to the library that provides it.
implemented in
Links an implementation to the library that provides it.
Links an MLIP run to the trained model it produces.
produces
Links an MLIP run to the trained model it produces.
Links an MLIP run to the training dataset it uses.
runs on
Links an MLIP run to the training dataset it uses.
Links an MLIP method to the simulation types it supports.
supports simulation
Links an MLIP method to the simulation types it supports.
Shortcut: links a trained model directly to the training dataset it was fitted on.
trained on
Shortcut: links a trained model directly to the training dataset it was fitted on.
]]>Shortcut: links a trained model directly to the MLIP method it applies.
trained with
Shortcut: links a trained model directly to the MLIP method it applies.
]]>Shortcut: links a trained model directly to a hyperparameter setting used during the MLIP run that produced it.
trained using
Shortcut: links a trained model directly to a hyperparameter setting used during the MLIP run that produced it.
]]>Links an MLIP method to the atomic environment descriptor it uses to represent local neighborhoods.
has descriptor
Links an MLIP method to the atomic environment descriptor it uses.
Type: AtomicEnvironmentDescriptor
Moment Tensor Descriptor
Invariant moment tensors of the neighborhood; used by MTP.
]]>Type: AtomicEnvironmentDescriptor
SOAP Descriptor
Smooth Overlap of Atomic Positions; used by GAP.
]]>Type: AtomicEnvironmentDescriptor
Symmetry Function Descriptor
Behler-Parrinello symmetry functions; used by HDNNP.
]]>Type: AtomicEnvironmentDescriptor
Equivariant Message Passing Descriptor
Equivariant message-passing representation; used by MACE and NequIP.
]]>Type: AtomicEnvironmentDescriptor
Atomic Cluster Expansion Descriptor
Atomic cluster expansion; used by ACE.
]]>Cutoff radius for atomic interactions, in angstroms.
xsd:double
cutoff radius
Cutoff radius for atomic interactions, in angstroms.
rdfs:Literal
default value
The expected datatype of this hyperparameter (e.g., float, int, string).
xsd:string
hyperparameter datatype
The expected datatype of this hyperparameter (e.g., float, int, string).
xsd:string
hyperparameter name
Learning rate used during model training.
xsd:double
learning rate
Learning rate used during model training.
rdfs:Literal
maximum value
rdfs:Literal
minimum value
Number of angular basis functions in the descriptor.
xsd:integer
number of angular basis functions
Number of angular basis functions in the descriptor.
Number of layers in the neural network architecture.
xsd:integer
number of layers
Number of layers in the neural network architecture.
Number of radial basis functions in the descriptor.
xsd:integer
number of radial basis functions
Number of radial basis functions in the descriptor.
rdfs:Literal
setting value
xsd:string
version
Asymptotic computational complexity of fitting the MLIP method, as a prior-knowledge annotation (e.g., "O(N)", "O(N log N)"). This is a property of the method itself, independent of any specific MLIP run.
xsd:string
training complexity
Asymptotic computational complexity of training.
Asymptotic computational complexity of running the trained potential on a system, as a prior-knowledge annotation (e.g., "O(N)" for local potentials).
xsd:string
inference complexity
Asymptotic computational complexity of inference.
Whether the algorithm (in its canonical implementation) supports GPU acceleration.
xsd:boolean
supports GPU
Whether the algorithm supports multi-node or multi-core parallel training and inference.
xsd:boolean
supports parallelization
Measured wall-clock time of a training run, in hours.
xsd:double
training duration
Measured wall-clock training time (hours).
Hardware used in the MLIP run (e.g., "NVIDIA A100", "Intel Xeon 8-core CPU").
xsd:string
training hardware
Peak memory usage during training, in gigabytes.
xsd:double
peak memory
Peak memory during training (GB).
Total GPU-hours consumed by the training run.
xsd:double
GPU hours
A single atomic configuration (positions, cell, properties) in a training dataset.
Subclass of: cmso:AtomicStructure
Atomic Configuration
A single atomic configuration (positions, cell,
properties) in a training dataset.
A physical property covered by the training dataset (energy, forces, stresses, virials).
Covered Property
A physical property covered by the training dataset
(energy, forces, stresses, virials).
]]>A density functional theory calculation that produced reference data for training.
Subclass of: mlips:ReferenceCalculation, mdo:Calculation
DFT Calculation
A density functional theory calculation that produced
reference data for training.
]]>Settings of a DFT calculation: exchange-correlation functional, k-point mesh, energy cutoff, pseudopotential type.
Subclass of: mlips:ReferenceSettings
DFT Settings
Settings of a DFT calculation: exchange-correlation
functional, k-point mesh, energy cutoff, pseudopotential type.
The provenance type of a training dataset.
Dataset Provenance
The provenance type of a training dataset.
]]>The basis set used in a DFT calculation, when the code is all-electron or otherwise uses a localised basis (e.g., FHI-aims NAO families, LAPW, Gaussian-type orbitals such as cc-pVDZ or 6-31G). Range of $\dlRole{dftBasisSet}$ on $\DFTSettingsC$. Plane- wave codes specify the basis through $\dlRole{energyCutoff}$ instead and leave this slot empty. Concrete basis sets are modelled as named individuals; an open extension point. Where the corpus uses a basis set not in the controlled vocabulary, a paper-local instance is minted and tagged with $\dlRole{candidateForVocabulary}$ for later review.
DFT Basis Set
The basis set used in a DFT calculation,
when the code is all-electron or otherwise uses a localised basis (e.g.,
FHI-aims NAO families, LAPW, Gaussian-type orbitals such as cc-pVDZ
or 6-31G). Range of mlips:dftBasisSet on DFTSettings. Plane-wave
codes specify the basis through mlips:energyCutoff instead and leave
this slot empty. Concrete basis sets are modelled as named individuals;
an open extension point.
]]>A material system studied in a dataset: element, alloy, compound, or complex microstructure.
Subclass of: cmso:CrystallineMaterial
Material System
A material system studied in a dataset: element, alloy,
compound, or complex microstructure.
A family of pseudopotentials used in a plane-wave DFT calculation (e.g., PAW, ultrasoft, norm-conserving). Range of $\dlRole{pseudopotentialType}$ on $\DFTSettingsC$. All-electron codes leave this slot empty. Concrete pseudopotential families are modelled as named individuals; an open extension point.
Pseudopotential Type
A family of pseudopotentials used in a
plane-wave DFT calculation (e.g., PAW, ultrasoft, norm-conserving).
Range of mlips:pseudopotentialType on DFTSettings. All-electron codes
leave this slot empty. Concrete pseudopotential families are modelled
as named individuals; an open extension point.
]]>A first-principles calculation that produced reference energies, forces, or stresses for MLIP training. Generalisation of DFTCalculation that also covers wave-function methods (e.g., CCSD(T), MP2, CASPT2) and other ab initio approaches.
Subclass of: mdo:Calculation
Reference Calculation
A first-principles calculation that produced
reference energies, forces, or stresses for MLIP training. Generalisation of
DFTCalculation that also covers wave-function methods (e.g., CCSD(T), MP2,
CASPT2) and other ab initio approaches.
]]>Settings of a reference calculation, abstracted across method families. Concrete subclasses (DFTSettings, WaveFunctionSettings) carry the method-specific parameters; the supertype carries cross-family slots such as the software code used.
Reference Settings
Settings of a reference calculation, abstracted
across method families. Concrete subclasses (DFTSettings, WaveFunctionSettings)
carry the method-specific parameters; the supertype carries cross-family slots
such as the software code used.
]]>A strategy used to sample the atomic configurations in a training dataset (e.g., chemical sampling across compositions, vibrational sampling across thermal snapshots, or a combination). Concrete sampling strategies are modeled as instances and will be developed in future work; this class is left open as an extension point.
Sampling Strategy
A strategy used to sample the atomic
configurations in a training dataset (e.g., chemical sampling across
compositions, vibrational sampling across thermal snapshots, or a
combination). Concrete sampling strategies are modeled as instances and
will be developed in future work; this class is left open as an
extension point.
]]>A dataset of atomic configurations used for MLIP training, with provenance, size, property coverage, and DFT settings.
Subclass of: mlschema:Dataset, prov:Entity
Training Dataset
A dataset of atomic configurations used for MLIP
training, with provenance, size, property coverage, and DFT settings.
]]>A wave-function-based reference calculation (e.g., Hartree--Fock, MP2, CCSD, CCSD(T), CASPT2). An alternative to DFTCalculation as a source of reference data, typically more accurate but more expensive.
Subclass of: mlips:ReferenceCalculation
Wave Function Calculation
A wave-function-based reference calculation
(e.g., Hartree-Fock, MP2, CCSD, CCSD(T), CASPT2). An alternative to
DFTCalculation as a source of reference data, typically more accurate but
more expensive.
]]>Settings of a wave-function-based reference calculation: the method (e.g., CCSD(T), MP2), the basis set (e.g., cc-pVTZ, aug-cc-pVDZ), and frozen-core treatment.
Subclass of: mlips:ReferenceSettings
Wave Function Settings
Settings of a wave-function-based reference
calculation: the method (e.g., CCSD(T), MP2), the basis set (e.g.,
cc-pVTZ, aug-cc-pVDZ), and frozen-core treatment.
A wave-function correlation method used to produce reference data (e.g., HF, MP2, CCSD, CCSD(T), DLPNO-CCSD(T), CASPT2, NEVPT2). Range of $\dlRole{wfMethod}$ on $\WaveFunctionSettingsC$. Concrete methods are modelled as named individuals; an open extension point.
Wave Function Method
A wave-function correlation method used to
produce reference data (e.g., HF, MP2, CCSD, CCSD(T), DLPNO-CCSD(T),
CASPT2, NEVPT2). Range of mlips:wfMethod on WaveFunctionSettings.
Concrete methods are modelled as named individuals; an open
extension point.
]]>A density-functional approximation to the exchange-correlation energy used in a DFT calculation (e.g., LDA, PBE, PBE0, HSE06, SCAN, BLYP, omegaB97X). Range of $\dlRole{xcFunctional}$ on $\DFTSettingsC$. Concrete functionals are modelled as named individuals; an open extension point.
Exchange-Correlation Functional
A density-functional approximation to the
exchange-correlation energy used in a DFT calculation (e.g., LDA, PBE,
PBE0, HSE06, SCAN, BLYP, omegaB97X). Range of mlips:xcFunctional on
DFTSettings. Concrete functionals are modelled as named individuals;
an open extension point.
]]>Annotates a paper-local instance with the controlled-vocabulary class to which it should be considered for promotion. Used when an encoded paper introduces an entity (e.g., a new exchange-correlation functional, a new wave-function method, a new pseudopotential family) that is not yet a named individual in mlips-vocab.ttl. The deduplication audit collects these candidates; a curator promotes, merges, or leaves them paper-local. Range is the target vocabulary class (e.g., $\XCFunctionalC$, $\WfMethodC$, $\DftBasisSetC$, $\PseudopotentialTypeC$).
owl:Thingowl:Class
candidate for vocabulary
Annotates a paper-local instance with the
controlled-vocabulary class to which it should be considered for
promotion. Used when an encoded paper introduces an entity that is not
yet a named individual in mlips-vocab.ttl. The deduplication audit
collects these candidates; a curator promotes, merges, or leaves them
paper-local.
]]>
covers material
covers property
dataset provenance
The basis set used by an all-electron or localised-basis DFT calculation (FHI-aims NAO, LAPW, Gaussian-type orbitals such as cc-pVDZ or 6-31G). Plane-wave codes specify the basis through $\dlRole{energyCutoff}$ and leave this slot empty.
DFT basis set
The basis set used by an all-electron or
localised-basis DFT calculation (FHI-aims NAO, LAPW, Gaussian-type
orbitals such as cc-pVDZ or 6-31G). Plane-wave codes specify the basis
through mlips:energyCutoff and leave this slot empty.
has configuration
Subproperty of: mlips:hasReferenceCalculation
has DFT calculation
Subproperty of: mlips:hasReferenceSettings
has DFT settings
Links a training dataset to the first-principles calculation that produced its reference data. Generalises hasDFTCalculation to wave-function and other ab initio methods.
has reference calculation
Links a training dataset to the
first-principles calculation that produced its reference data.
Generalises hasDFTCalculation to wave-function and other ab initio
methods.
Links a reference calculation to its method-specific settings. Generalises hasDFTSettings to wave-function and other ab initio methods.
has reference settings
Links a reference calculation to its
method-specific settings. Generalises hasDFTSettings to wave-function
and other ab initio methods.
Subproperty of: mlips:hasReferenceSettings
has wave function settings
The family of pseudopotentials used in a plane-wave DFT calculation (PAW, ultrasoft, norm-conserving). All- electron codes leave this slot empty and use $\dlRole{dftBasisSet}$ instead.
pseudopotential type
The family of pseudopotentials used in a
plane-wave DFT calculation (PAW, ultrasoft, norm-conserving). All-
electron codes leave this slot empty and use mlips:dftBasisSet
instead.
Records the sampling strategy or strategies used to construct a training dataset. Optional and multi-valued: a dataset may combine several sampling strategies (e.g., chemical and vibrational).
sampling strategy
Records the sampling strategy or strategies
used to construct a training dataset. Optional and multi-valued: a
dataset may combine several sampling strategies (e.g., chemical and
vibrational).
The DFT software used (e.g., VASP, GPAW).
Subproperty of: mlips:usedReferenceCode
used DFT code
The DFT software used (e.g., VASP, GPAW).
The software code used to perform the reference calculation (e.g., VASP, GPAW for DFT; Molpro, PySCF, ORCA for wave-function methods). Generalises usedDFTCode.
used reference code
The software code used to perform the
reference calculation (e.g., VASP, GPAW for DFT; Molpro, PySCF, ORCA
for wave-function methods). Generalises usedDFTCode.
The wave-function correlation method used to produce reference data (e.g., HF, MP2, CCSD, CCSD(T), DLPNO-CCSD(T), CASPT2, NEVPT2). Range is $\WfMethodC$; concrete methods are named individuals in mlips-vocab.ttl.
wave-function method
The wave-function correlation method used
to produce reference data (e.g., HF, MP2, CCSD, CCSD(T), DLPNO-CCSD(T),
CASPT2, NEVPT2). Range is WfMethod; concrete methods are named
individuals in mlips-vocab.ttl.
The exchange-correlation functional used in a DFT calculation (e.g., LDA, PBE, PBE0, HSE06, SCAN, omegaB97X). Range is $\XCFunctionalC$; concrete functionals are named individuals in mlips-vocab.ttl.
exchange-correlation functional
The exchange-correlation functional used in
a DFT calculation (e.g., LDA, PBE, PBE0, HSE06, SCAN, omegaB97X). Range
is XCFunctional; concrete functionals are named individuals in
mlips-vocab.ttl.
The Gaussian basis set used in a wave-function calculation, e.g., cc-pVTZ, aug-cc-pVDZ, def2-TZVP.
xsd:string
basis set
The Gaussian basis set used in a
wave-function calculation, e.g., cc-pVTZ, aug-cc-pVDZ, def2-TZVP.
xsd:string
chemical formula
Plane-wave energy cutoff in eV.
xsd:double
energy cutoff
Plane-wave energy cutoff in eV.
Whether core electrons are kept frozen during the post-Hartree--Fock correlation treatment in a wave-function calculation. True for the standard frozen-core approximation; false for all-electron correlation.
xsd:boolean
frozen core
Whether core electrons are kept frozen
during the post-Hartree-Fock correlation treatment in a wave-function
calculation. True for the standard frozen-core approximation; false
for all-electron correlation.
K-point mesh specification (e.g., '4x4x4').
xsd:string
k-point mesh
K-point mesh specification (e.g., '4x4x4').
E.g., element, binary alloy, ternary compound, HEA.
xsd:string
material class
E.g., element, binary alloy, ternary compound, HEA.
E.g., point defect, dislocation, surface, grain boundary.
xsd:string
microstructural feature
E.g., point defect, dislocation, surface, grain boundary.
xsd:integer
number of configurations
Dataset augmented from existing sources.
Type: DatasetProvenance
Augmented
Dataset augmented from existing sources.
]]>Type: CoveredProperty
Energy
]]>Type: CoveredProperty
Forces
]]>Dataset generated by in-house calculations.
Type: DatasetProvenance
In-house
Dataset generated by in-house calculations.
]]>Dataset from a published source.
Type: DatasetProvenance
Published
Dataset from a published source.
]]>Type: CoveredProperty
Stresses
]]>Type: CoveredProperty
Virials
]]>An accuracy measure reported in a benchmark (e.g., RMSE of energy, MAE of forces).
Accuracy Metric
An accuracy measure reported in a benchmark
(e.g., RMSE of energy, MAE of forces).
1
1
]]>A single evaluation result: an MLIP method applied to a material system with specific hyperparameters, reporting accuracy metrics.
Benchmark Result
A single evaluation result: an MLIP method applied to a
material system with specific hyperparameters, reporting accuracy metrics.
1
]]>A published study reporting MLIP evaluation results.
Subclass of: prov:Activity
Benchmark Study
A published study reporting MLIP evaluation results.
]]>The physical property being measured (energy, force, stress).
Metric Property
The physical property being measured (energy, force, stress).
]]>Type of accuracy metric (RMSE, MAE, R², etc.).
Metric Type
Type of accuracy metric (RMSE, MAE, R², etc.).
]]>
has accuracy metric
Links an MLIP run to a concrete value chosen for one of the hyperparameters of the MLIP method it applies. The collection of these settings, together with the method and the training dataset, fully characterises what was fitted in the run.
has hyperparameter setting
Links an MLIP run to a concrete hyperparameter setting used in that run.
has result
metric property
metric type
reported in
target material
uses algorithm
uses training data
xsd:double
metric value
Measured wall-clock inference time per atom, in microseconds. Captured on a specific benchmark result.
xsd:double
inference time per atom
Inference time per atom (microseconds).
Hardware used for the inference benchmark (e.g., "NVIDIA A100", "Apple M2").
xsd:string
inference hardware
Type: MetricProperty
Energy (metric property)
]]>Type: MetricProperty
Force
]]>Mean Absolute Error.
Type: MetricType
MAE
Mean Absolute Error.
]]>Coefficient of determination.
Type: MetricType
R²
Coefficient of determination.
]]>Root Mean Square Error.
Type: MetricType
RMSE
Root Mean Square Error.
]]>Type: MetricProperty
Stress
]]>Links a numeric value to its QUDT unit.
has unit
Links a numeric value to its QUDT unit.
Links an entity to its corresponding Wikidata item.
same as Wikidata
Links an entity to its corresponding Wikidata item.