1. Final thoughts
Congratulations! Now you have mastered predicting churn using Decision Trees. One thing that I would like to note here, is that although we concentrated on this model, Decision Trees are not the only choice to predict employee turnover.
2. Alternative methods
One very popular alternative that is used widely in HR analytics is Logistic Regression. You may still use Python and sklearn to make predictions with Logistic Regression using the evaluation metrics and approaches discussed in this course.
While single Decision Tree is good, sometimes many is better. The tree based algorithms like Random Forest or Gradient Boosting usually provide better results than a single Decision Tree. The reason we use single one, is that those complex models are not interpretable and cannot be visualized to make decisions using Decision path.
Last but not list, Neural networks are popular alternative nowadays for many prediction tasks including turnover prediction. However, they are considered a black box and do not provide the clue behind their predictions, which especially in HR is very important.
As you have completed this course, now my advice would be to take some other HR datasets and attach your predictive skills on them following the same order of tasks that we have completed during this course.
3. The End