Articles
Identifying Potential High-Risk Factor Arrhythmia Patients Using Predictive Modeling
Top predictive features of arrhythmias from the diabetes patient model. Plots demonstrate the positive correlation between age and predictive importance. Plots demonstrate the positive correlation between months from initial diabetes diagnosis and predictive importance.
Scoring and Segmenting Key Opinion Leader Physicians With Innovative and Customizable Machine Learning Techniques
Five segments of healthcare providers based specifically on the likelihood of adopting the oncology product for first line use.
Improving an AATD Predictive Model Using EMR Data
SHAP values of top features from the highest-performing predictive model, including diagnoses, lab results and location of service factors.
Use of Machine Learning to Identify Gastroparesis Patients Suitable for Nasal Spray Metoclopramide
Poster presented at Digestive Disease Week 2023 outlining model performance, feature importance and patient clustering in gastroparesis patients suitable for nasal spray metoclopramide.
Predictive Modeling for Treatment Switching in Paroxysmal Nocturnal Hemoglobinuria (PNH) Patients
Plots comparing feature importance between the patient-based model, the physician-based model and the hybrid patient-physician-based model.