ICD10PCS Entity Resolver

Description

Entity Resolution model Based on KNN using Word Embeddings + Word Movers Distance.

Predicted Entities

ICD10-PCS Codes and their normalized definition with clinical_embeddings.

Download

How to use

...
model = ChunkEntityResolverModel.pretrained("chunkresolve_icd10pcs_clinical","en","clinical/models")
	.setInputCols("token","chunk_embeddings")
	.setOutputCol("entity")
    
pipeline_icd10pcs = Pipeline(stages = [documentAssembler, sentenceDetector, tokenizer, stopwords, word_embeddings, ner, chunk_embeddings, model])

data = ["""He has a starvation ketosis but nothing found for significant for dry oral mucosa"""]

pipeline_model = pipeline_icd10pcs.fit(spark.createDataFrame([[""]]).toDF("text"))

light_pipeline = LightPipeline(pipeline_model)

result = light_pipeline.annotate(data)
...
val model = ChunkEntityResolverModel.pretrained("chunkresolve_icd10pcs_clinical","en","clinical/models")
	.setInputCols("token","chunk_embeddings")
	.setOutputCol("entity")
    
val pipeline = new Pipeline().setStages(Array(documentAssembler, sentenceDetector, tokenizer, stopwords, word_embeddings, ner, chunk_embeddings, model))

val result = pipeline.fit(Seq.empty["He has a starvation ketosis but nothing found for significant for dry oral mucosa"].toDS.toDF("text")).transform(data)

Results

|   | chunks               | begin | end | code    | resolutions                                      |
|---|----------------------|-------|-----|---------|--------------------------------------------------|
| 0 | a starvation ketosis | 7     | 26  | 6A3Z1ZZ | Hyperthermia, Multiple:::Narcosynthesis:::Hype...|
| 1 | dry oral mucosa      | 66    | 80  | 8E0ZXY4 | Yoga Therapy:::Release Cecum, Open Approach:::...|

Model Information

Name: chunkresolve_icd10pcs_clinical  
Type: ChunkEntityResolverModel  
Compatibility: Spark NLP 2.4.2+  
License: Licensed  
Edition: Official  
Input labels: token, chunk_embeddings  
Output labels: entity  
Language: en  
Case sensitive: True  
Dependencies: embeddings_clinical  

Data Source

Trained on ICD10 Procedure Coding System dataset https://www.icd10data.com/ICD10PCS/Codes