Project Results
A deeper look at the outcome of my work.
To predict recurrence I used two different machine learning techniques and compared the accuracy of them. The first model was a Logistic Regression model predicting recurrence, it had a 93.5% accuracy, which was more than satisfactory. This value did not change with attempted hyperparameter tuning. The second model was a Decision Tree model which was slightly less accurate with a percentage of 89.6%. I raised this value up to 92.2%, hypertuning the maximum depth, which maximized at 5. While both of these models do a great job of predicting recurrence, the Logistic Regression model was slightly more accurate at predicting cancers return.