Decision Tree

Introduction In the previous article, we talked about how to use Bagging to build a random forest. To recap, random forest combines multiple independent decision trees and makes the final classification based on the majority vote. However, we can do more — we can build a set of dependent decision…

Understanding CART — CART is an algorithm that uses decision trees to solve classification and regression problems. The core algorithm for building decision tree is called ID3 (Iterative Dichotomiser 3). ID3 (Iterative Dichotomiser 3) The ID3 algorithm begins with the original dataset as the root node. On each iteration of the algorithm, it iterates through every unused…

Introduction In the previous article, we talked about decision tree in terms of its structure and parameters, as well as the example of Titanic dataset. However, decision tree is essentially a weak learner since the accuracy of a tree model is generally low (i.e., …

Decision Tree is a Supervised learning technique that can be used for both classification and Regression problems, but mostly it is preferred for solving Classification problems. …

In this article, we’ll discuss how to choose the class of an impure leaf node in a decision tree. For simplicity, we’ll use the titanic dataset from Kaggle. Let’s look at the training dataset. Here is a distribution of whether the passengers survived or perished:

Decision Trees are one of the best known supervised classification methods. “A decision tree is a way of representing knowledge obtained in the inductive learning process. The space is split using a set of conditions, and the resulting structure is the tree“ A tree is composed of nodes, and those…

Machine learning is self-learning algorithm whose performance improves as it is exposed to more data over time. Machine learning is a advanced progressive concept in dealing with complex tasks which cannot be handled by user-programmed code alone. It is widely used in enterprises and even small companies for the purpose…

One thing to know is that machines do not do well with text. They prefer to work with numbers. Therefore, anytime we come across a feature/variable column in our dataset that contains text, we need to convert this data type to numbers in order for the algorithm to process them…