Pipeline for Anatomic Therapeutic Chemical (ATC) Sentence Entity Resolver

Description

This advanced pipeline extracts DRUG entities from clinical texts and utilizes the sbiobert_base_cased_mli Sentence Bert Embeddings to map these entities to their corresponding Anatomic Therapeutic Chemical (ATC) codes.

Predicted Entities

DRUG

Live Demo Open in Colab Download Copy S3 URI

How to use

from sparknlp.pretrained import PretrainedPipeline

ner_pipeline = PretrainedPipeline("atc_resolver_pipeline", "en", "clinical/models")

result = ner_pipeline.annotate("""She was immediately given hydrogen peroxide 30 mg and amoxicillin twice daily for 10 days to treat the infection on her leg. She has a history of taking magnesium hydroxide.""")

import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline

val ner_pipeline = PretrainedPipeline("atc_resolver_pipeline", "en", "clinical/models")

val result = ner_pipeline.annotate("""She was immediately given hydrogen peroxide 30 mg and amoxicillin twice daily for 10 days to treat the infection on her leg. She has a history of taking magnesium hydroxide.""")

Results

|    | chunks              |   begin |   end | entities   | atc_code   | resolutions         |
|---:|:--------------------|--------:|------:|:-----------|:-----------|:--------------------|
|  0 | hydrogen peroxide   |      26 |    42 | DRUG       | A01AB02    | hydrogen peroxide   |
|  1 | amoxicillin         |      54 |    64 | DRUG       | J01CA04    | amoxicillin         |
|  2 | magnesium hydroxide |     153 |   171 | DRUG       | A02AA04    | magnesium hydroxide |

Model Information

Included Models

• DocumentAssembler
• SentenceDetectorDLModel
• TokenizerModel
• WordEmbeddingsModel
• MedicalNerModel
• NerConverterInternalModel
• Chunk2Doc
• BertSentenceEmbeddings
• SentenceEntityResolverModel