Description
This advanced pipeline extracts medication entities from clinical texts and utilizes the sbiobert_base_cased_mli Sentence Bert Embeddings to map these entities to their corresponding National Drug Codes (NDC) codes.
Predicted Entities
NDC
How to use
from sparknlp.pretrained import PretrainedPipeline
ndc_pipeline = PretrainedPipeline("ndc_resolver_pipeline", "en", "clinical/models")
result = ndc_pipeline.fullAnnotate("""The patient was given aspirin 81 mg and metformin 500 mg""")
import com.johnsnowlabs.nlp.pretrained.PretrainedPipeline
val ndc_pipeline = PretrainedPipeline("ndc_resolver_pipeline", "en", "clinical/models")
val result = ndc_pipeline.fullAnnotate("""The patient was given aspirin 81 mg and metformin 500 mg""")
Results
+----------------+-----+---+---------+----------+----------------+------------------------------------------------------------+
| chunk|begin|end|ner_label| ndc_code| description| aux_labels|
+----------------+-----+---+---------+----------+----------------+------------------------------------------------------------+
| aspirin 81 mg| 22| 34| DRUG|41250-0780| aspirin 81 mg|{'packages': "['1 BOTTLE, PLASTIC in 1 PACKAGE (41250-780...|
|metformin 500 mg| 40| 55| DRUG|62207-0491|metformin 500 mg|{'packages': "['5000 TABLET in 1 POUCH (62207-491-31)', '...|
+----------------+-----+---+---------+----------+----------------+------------------------------------------------------------+
Model Information
| Model Name: | ndc_resolver_pipeline |
| Type: | pipeline |
| Compatibility: | Healthcare NLP 5.4.1+ |
| License: | Licensed |
| Edition: | Official |
| Language: | en |
| Size: | 2.9 GB |
Included Models
- DocumentAssembler
- SentenceDetectorDLModel
- TokenizerModel
- WordEmbeddingsModel
- TextMatcherInternalModel
- MedicalNerModel
- NerConverter
- MedicalNerModel
- NerConverterInternalModel
- ChunkMergeModel
- Chunk2Doc
- BertSentenceEmbeddings
- SentenceEntityResolverModel