CART (Classification And Regression Tree) in Machine Learning

In this video, we will learn about CART (Classification And Regression Tree) algorithm in Machine Learning. 

CART algorithm works very similar to the decision trees but are little bit more sophisticated. There is a term known as Gini index which is the core part of the CART algorithm. Gini index is calculated by subtracting the squared summation of probability of occurrence of each class from one. 

This Gini index helps us to determine that which feature will be present at which level of the tree structure as there are 3 types of nodes in a tree. 
- Root Nodes - This is the node from which the tree start. 
- Internal Nodes - They are the internal nodes which makes the tree highly dense and sometimes leads to overfitting. 
- Leaf Nodes - They are the final predictors for the target class. 

Some of the advantages to use CART algorithm are robustness to outliers and are helps to determine non-linear boundaries between the different target classes present in the dataset. But due to high complexity sometimes it leads to overfitting and leads to unstable tree structure. 
 

CART in Machine Learning:
https://www.geeksforgeeks.org/cart-classification-and-regression-tree-in-machine-learning/