Analysis of readability and structural accuracy in SNOMED CT - Supplemetary material

This repository contains the supplementary material for the article Analysis of readability and structural accuracy in SNOMED CT. The files are described next.

Additional file 1 --- snomed_history_metrics.tsv

Tabular file containing the value of all metrics along all the SNOMED CT versions considered in the article. The file contains the following columns:

• File: The ontology file in which the metric is calculated. The name of the files is in the format YYYYMMDD (year, month and day), indicating the SNOMED CT version.
• Metric: The name of the metric.
• Value: The metric value for the corresponding SNOMED CT version.

Additional file 2 --- detailed_files_snomed2019

Folder containing an individual tabular file per metric applied to SNOMED CT 20190731. Each file contains concrete information for each metric. In the case of readability metrics, the columns are the following:

• Metric: Metric name.
• Entity: The individual entity that is measured by the metric. Depending on the metric, this column will be named "Class", "Object property", "Data Property", or "Annotation Property".
• Metric Value: The individual metric value for the entity.
• Hierarchy: The SNOMED CT hierarchy in which the entity is located. This value is only present if the entity is a class.

In the case of structural accuracy metrics, the columns are the following:

• Metric: Metric name.
• Class: The LR class that is measured.
• Class depth: The depth in which the LR class is located in the ontology.
• LR: The label of the LR class, which is a lexical regularity.
• Positive Cases: Number of positive cases. If metric is "Lexically suggest logically define", the value of this column is the number of classes exhibiting LR in their labels that are semantically related to Class. If metric is "Systematic naming", positive cases are the number of subclasses of Class that exhibit LR.
• Positive cases average depth: The average depth in which the positive cases are located in the ontology.
• Positive cases average distance to LR class: The average hierarchical distance from positive cases to Class.
• Negative Cases: Number of negative cases. If metric is "Lexically suggest logically define", the value of this column is the number of classes exhibiting LR in their labels that are not semantically related to Class. If metric is "Systematic naming", positive cases are the number of subclasses of Class that do not exhibit LR.
• Negative cases average depth: The average depth in which the negative cases are located in the ontology.
• Negative cases average distance to LR class: The average hierarchical distance from negative cases to Class.
• Metric Value: The individual metric value for the LR class according the formula $\frac{Positive Cases}{Positive Cases + NegativeCases}$.
• Hierarchy: The SNOMED CT hierarchy in which the class is located.

Additional file 3 --- snomed_last_version_lsld.log

Log file including information of the negative cases detected in SNOMED CT 20190731 according to LSLD metric.

Additional file 4 --- snomed_last_version_systematic_naming.log

Log file including information of the negative cases detected in SNOMED CT 20190731 according to systematic naming metric.