2.3.3.3.16. NXapm_compositionspace_config

Status:

application definition, extends NXobject

Description:

Application definition for a configuration of the CompositionSpace tool used in ...

Application definition for a configuration of the CompositionSpace tool used in atom probe.

This is an application definition for the common NFDI-MatWerk/FAIRmat infrastructure use case IUC09 that explores how to improve the organization and results storage of the CompositionSpace software using NeXus.

Symbols:

No symbol table

Groups cited:

NXentry, NXidentifier, NXobject, NXprocess, NXserialized

Structure:

ENTRY: (required) NXentry

definition: (required) NX_CHAR ⤆

Obligatory value: NXapm_compositionspace_config

@version: (optional) NX_CHAR ⤆

config: (required) NXobject

analysis_identifier: (recommended) NX_UINT

IDENTIFIER: (optional) NXidentifier

reconstruction: (required) NXserialized

Specification of the tomographic reconstruction used for this analysis. ...

Specification of the tomographic reconstruction used for this analysis. Typically, reconstructions in the field of atom probe tomography are communicated via files which store at least reconstructed ion positions and mass-to-charge-state-ratio values. Container files like HDF5 though can store multiple reconstructions. Therefore, the position and mass_to_charge concepts point to specific instances to use for this analysis.

type: (optional) NX_CHAR ⤆

path: (required) NX_CHAR ⤆

checksum: (recommended) NX_CHAR ⤆

algorithm: (recommended) NX_CHAR ⤆

position: (required) NX_CHAR

Name of the node which resolves the reconstructed ...

Name of the node which resolves the reconstructed ion position values to use for this analysis.

mass_to_charge: (optional) NX_CHAR

Name of the node which resolves the mass-to-charge-state ratio ...

Name of the node which resolves the mass-to-charge-state ratio values for each reconstructed ion to use for this analysis.

ranging: (required) NXserialized

Specification of the ranging definitions used for this analysis. ...

Specification of the ranging definitions used for this analysis.

Indices start from 1. The value 0 is reserved for the null model of unranged positions whose iontype is unknown_type. The value 0 is also reserved for voxels that lie outside the dataset.

type: (optional) NX_CHAR ⤆

path: (required) NX_CHAR ⤆

checksum: (recommended) NX_CHAR ⤆

algorithm: (recommended) NX_CHAR ⤆

ranging_definitions: (required) NX_CHAR

Name of the (parent) node directly below which the ranging definitions f ...

Name of the (parent) node directly below which the ranging definitions for (molecular) ions are stored.

voxelization: (required) NXprocess

Step during which the point cloud is discretized to compute element-specif ...

Step during which the point cloud is discretized to compute element-specific composition fields. Iontypes are atomically decomposed to correctly account for the multiplicity of each element that was ranged for each ion.

edge_length: (required) NX_NUMBER {units=NX_LENGTH}

Edge length of cubic voxels building the 3D grid that is used for discre ...

Edge length of cubic voxels building the 3D grid that is used for discretizing the point cloud.

autophase: (required) NXprocess

Optional step during which the subsequent segmentation step is prepared wi ...

Optional step during which the subsequent segmentation step is prepared with the aim to eventually reduce the dimensionality of the chemical space in which the machine learning model works.

In this step a supervised reduction of the dimensionality of the chemical space is quantified using the (Gini) feature importance of each element to suggest which columns of the composition matrix should be taken for the subsequent segmentation step.

use: (required) NX_BOOLEAN

Was the automated phase assignment used?

initial_guess: (required) NX_POSINT {units=NX_UNITLESS}

Estimated guess for which a Gaussian mixture model is evaluted to prepro ...

Estimated guess for which a Gaussian mixture model is evaluted to preprocess a result that is subsequenty post-processed with a random_forest_classifier to lower the number of dimensions in the chemical space to the subset of trunc_species many elements with the highest feature importance.

trunc_species: (required) NX_POSINT {units=NX_UNITLESS}

The number of elements to use for reducing the dimensionality.

random_forest_classifier: (optional) NXprocess

Configuration for the random forest classification model.

segmentation: (required) NXprocess

Step during which the voxel set is segmented into voxel sets with differen ...

Step during which the voxel set is segmented into voxel sets with different chemical composition.

pca: (optional) NXprocess

A principal component analysis of the chemical space to guide a decision ...

A principal component analysis of the chemical space to guide a decision into how many sets of voxels with different chemical composition the machine learning algorithm suggests to split the voxel set.

ic_opt: (required) NXprocess

The decision is guided through the evalution of the information criterio ...

The decision is guided through the evalution of the information criterion minimization.

n_max_ic_cluster: (required) NX_POSINT {units=NX_UNITLESS}

The maximum number of chemical classes to probe with the Gaussian mixt ...

The maximum number of chemical classes to probe with the Gaussian mixture model with which the voxel set is segmented into a mixture of voxels with that many different chemical compositions.

gaussian_mixture: (optional) NXprocess

Configuration for the Gaussian mixture model that is used in the segme ...

Configuration for the Gaussian mixture model that is used in the segmentation step.

clustering: (required) NXprocess

Step during which the chemically segmented voxel sets are analyzed for the ...

Step during which the chemically segmented voxel sets are analyzed for their spatial organization.

dbscan: (required) NXobject

Configuration for the DBScan algorithm that is used in the clustering step.

eps: (required) NX_FLOAT {units=NX_LENGTH}

The maximum distance between voxel pairs in a neighborhood to be consi ...

The maximum distance between voxel pairs in a neighborhood to be considered connected.

min_samples: (required) NX_UINT {units=NX_UNITLESS}

The number of voxels in a neighborhood for a voxel to be considered as ...

The number of voxels in a neighborhood for a voxel to be considered as a core point.

Hypertext Anchors

List of hypertext anchors for all groups, fields, attributes, and links defined in this class.

NXDL Source:

https://github.com/FAIRmat-NFDI/nexus_definitions/tree/fairmat/contributed_definitions/NXapm_compositionspace_config.nxdl.xml