Machine Learning & NLP for Decision Support
A radiology firm required a means to more quickly and accurately provide best practice recommendations based on dictated reports describing medical images.
Illumination Works created a rapid algorithm to analyze dictated report text in real time to provide radiologists with follow-up recommendations for identified conditions.
- Logical representations of best practice
- Annotated data sets that yield training sets
and truth keys
- Productionized, algorithmic software that applies the correct condition model
- Real-time follow-up care recommendations
output to custom user interface
- Log files to assess and retrain model performance
- Modular, object-oriented structure, scalable
for additional conditions
- Natural Language Processing (NLP)
- Machine learning Recommender System-based models
- Diagnosis decision trees