-
Notifications
You must be signed in to change notification settings - Fork 2
Home
The Information Extraction (IE) models provide a framework for explicitly defining phenotypes at the level of the patient, the encounter, a document, and a mention of a clinical concept in text. A user can create classes inside a customized domain IE model to engineer a computable phenotype definition that describes both the annotation variables comprising the phenotype and the linguistic information needed to identify those variables in text. Therefore, a domain information extraction model developed by a user also serves as a schema and a knowledge base for NLP tools. The IE models support development of NLP knowledge bases from domain experts and allow sharing and crowdsourcing of knowledge.
The files in the repository are the information extraction models created by and used in the BLU Lab. All versions of the models are protected under the Creative Commons Attribution 4.0 International Public License.
To suggest any edits to the following models, please submit an Issue here.
The models can also be opened by using the following URIs in an ontology editor:
- Schema: http://blulab.chpc.utah.edu/ontologies/v2/Schema.owl
- Modifier: http://blulab.chpc.utah.edu/ontologies/v2/Modifier.owl
- ConText: http://blulab.chpc.utah.edu/ontologies/v2/ConText.owl
- TermMapping: http://blulab.chpc.utah.edu/ontologies/TermMapping.owl
- Example Pneumonia Domain: http://blulab.chpc.utah.edu/ontologies/examples/pneumonia.owl
TermMapping.owl: This file contains a list of annotation properties to hold preferredTerms, synonymns, misspellings, codes, alternateCodes, and regex. It is imported into all of the other files.
ConText.owl: This file contains a set of classes and object properties to specify the components of the ConText algorithm developed by the BLU Lab. The components consist of classes to describe the modifiers, termination terms, pseudo terms and actions needed for an NLP system to extract modifiers related to a potential anchor/target. This file is imported and used by both the Modifer.owl and Schema.owl files.
Modifier.owl: This file extends the ConText.owl file to include a hierarchy of modifier terms that are beneficial for biomedical NLP information extraction. Also, included are instances that contain the lexical cues for each modifier class.
Schema.owl: This file contains an NLP schema that models the types of annotations that can be created from NLP systems. It contains a hierarcy of semantic categories that can be associated with certain linguistic and semantic modifiers. This file is imported into a domain ontology to create domain specific variables for extraction.
pneumonia.owl: This file is an example domain ontology that is based on the CDC case definition for pneumonia.