Advanced Machine Learning for Predicting Hospital Readmissions in Diabetes Care
Machine Learning
Random Forest
XGBoost
Decision Trees
Deep Learning
Logistic Regression
Table of Contents
Motivation
We apply numerous ML algorithms to predict whether a patient will be readmitted to the hospital (within or after 30 days of discharge).
Exploratory Data Analysis
Python packages used in the project.
Dash Board visualizing the raw data.
Feature Engineering
We appplied different methods to select important features so it will reduce the computational time, the risk of overfitting and
complexity of interpretation.
- Recursive Feature Elimination (RFE)
- \( \chi^2 \)
- Univariate Feature Selection : ANOVA F-value
- Information Value (IV) and Weight of evidence (WoE)
- Correlation
- Threshold
- Bortua Algorithm
Important features (shown in green) selected by different methods.
Results
Permutation feature importance.
Impact
Developed a model that predicts if a diabetes patient re-admit to the hospital or not Sucessfully, achieving a 0.7 Recall Rate.