Five ML algorithms were used on the heart failure dataset to determine which type will be most accurate in predicting patient mortality.
KNN, SVM, Decision Tree Classifier, Random Forest Classifier and Gradient Boosting Classifier were used for the classification of the dataset.
All ML calculations and data manipulation were done in Python. Results were exported to Tableau for visualization.
| Algorithm: | Accuracy Score: |
|---|---|
| KNN | 0.62 |
| SVM | 0.76 |
| Decision Tree classifier | 0.77 |
| Random Forest classifier | 0.84 |
| Gradient Boosting classifier | 0.86 |
The real outcome cluster vs the KNN cluster. The clusters were plotted on two dimensions, age and days under care.
Red represents a deceased paitent wereas green represents a live one.
By roughly comapring the two you can see that the knn cluster is matching to 62% accuracy.
The four supervised ML algorithms' predictions on testing data compared to the real outcome.
In this plot the dots marks the patients that were selcted as the testing data out ot the 300 paitents.
A dark dot represents a deceased paitent wereas a light one represents a live paitent.
