Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/leebrian/aws-cm-test
Running synthetic chief complaints through AWS Comprehend Medical
https://github.com/leebrian/aws-cm-test
Last synced: 8 days ago
JSON representation
Running synthetic chief complaints through AWS Comprehend Medical
- Host: GitHub
- URL: https://github.com/leebrian/aws-cm-test
- Owner: leebrian
- License: apache-2.0
- Created: 2018-12-20T02:01:37.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2021-06-25T15:24:00.000Z (over 3 years ago)
- Last Synced: 2023-03-24T07:24:10.032Z (over 1 year ago)
- Language: Python
- Size: 8.4 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: readme.md
- License: LICENSE
Awesome Lists containing this project
README
# Test AWS Comprehend Medical
Running synthetic chief complaints through AWS Comprehend Medical.
Using Scott Lee's public data sets from
Question- how does CM diagnosis compare to process that generated these codes using the process described in [Scott's paper](https://www.nature.com/articles/s41746-018-0070-0)? Using the set of 42,364 synthetic chief complaints and HCUP CCS diagnosis codes.
## Results
Multiple different types of entity types found ['SYSTEM_ORGAN_SITE', 'DX_NAME', 'ACUITY', 'DIRECTION', 'ID', 'PROCEDURE_NAME', 'TEST_NAME', 'TREATMENT_NAME', 'PHONE_OR_FAX', 'GENERIC_NAME', 'ADDRESS', 'BRAND_NAME', 'DATE', 'NAME', 'PROFESSION']
Multiple different types of entity categories found ['ANATOMY', 'MEDICAL_CONDITION', 'PROTECTED_HEALTH_INFORMATION', 'TEST_TREATMENT_PROCEDURE', 'MEDICATION']
Multiple different types of trait names found ['SYMPTOM', 'DIAGNOSIS', 'SIGN', 'NEGATION']
Multiple different tuples of type, caegory, trait found [('DX_NAME', 'MEDICAL_CONDITION', 'SYMPTOM'), ('DX_NAME', 'MEDICAL_CONDITION', 'DIAGNOSIS'), ('DX_NAME', 'MEDICAL_CONDITION', 'SIGN'), ('DX_NAME', 'MEDICAL_CONDITION', 'NEGATION')]
## Requirements
Remember to [configure your local environment](https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-configure.html#cli-quick-configuration) with an aws access key, secret key, and default region. Entire exploration done using free tier (thanks AWS) and in the us-east-1 region, although I don't think that matters.
## References
[CCS Users Guide with descriptions of 259 diagnosis categories](https://www.hcup-us.ahrq.gov/toolssoftware/ccs/CCSUsersGuide.pdf)
[AWS CM Docs](https://docs.aws.amazon.com/comprehend/latest/dg/comprehend-medical.html)
[AWS CM API Docs](https://docs.aws.amazon.com/comprehend/latest/dg/API_Reference.html)
[AWS CM Python Boto3 Docs](https://boto3.amazonaws.com/v1/documentation/api/latest/reference/services/comprehendmedical.html)
## Screenshots
## Data Maps
```json
[
{
"data-source": "ML-Workshop-Synthetic-Chief-Complaints",
"data-owner": "https://gitlab.com/gte577z",
"data-custodian": "https://gitlab.com/gte577z",
"address": "https://gitlab.com/gte577z/ml-workshop/raw/master/data/text.csv",
"extract-date": "2018-12-20",
"format": "CSV",
"methods": "https://doi.org/10.1038/s41746-018-0070-0",
"fields": [
"text",
"diagnosis"
]
},
{
"data-source": "AWS-Medical-Comprehend-Assess-Chief-Complaints",
"data-owner": "OCIO",
"data-custodian": "OCIO",
"address": "https://github.com/leebrian/aws-cm-test/blob/master/dat/text-acm.csv",
"format": "CSV",
"methods": "https://github.com/leebrian/aws-cm-test",
"tags": [
"AWS",
"Comprehend Medical"
],
"fields": [
"text",
"diagnosis",
"aws-text",
"aws-score",
"aws-category",
"aws-category",
"aws-trait"
],
"upstream": "ML-Workshop-Synthetic-Chief-Complaints"
}
]
```