From the "20 Questions game" to medical diagnosis - see how Decision Trees select yes/no questions using (weighted) entropy and information gain, and use pruning for better generalization. Learn the difference between black-box and white-box models.
I also included an explainable, real-world example on the Pima Indians Diabetes data.
More of my materials on Decision Trees:
- interactive Live Script designed as an aid in understanding the Decision Tree fundamentals (accessible through MATLAB Online): https://mathworks.com/matlabcentral/fileexchange/180564-introduction-to-decision-tree
- article on the Decision Tree basics (in Polish): https://ont.com.pl/articles/DecisionTree
Music:
- https://pixabay.com/music/meditationspiritual-piano-dreamcloud-meditation-179215/
References:
1. https://kaggle.com/datasets/uciml/pima-indians-diabetes-database
2. https://en.wikipedia.org/wiki/Decision_tree
3. https://en.wikipedia.org/wiki/Decision_tree_learning
Timestamps:
00:00 Introduction
01:10 "What to wear" example
02:04 Decision Tree structure
03:45 Diabetes diagnosis example
06:55 Proper split
09:15 Entropy
12:58 Information gain
15:45 Next splits
17:22 Decision Tree interpretation
17:54 Pruning
18:28 New data samples usage
18:55 White box model
19:08 Timeline
19:28 Pros and cons
