Break down complex decisions into smaller pieces to be a better decision maker!
This is exactly what Decision Trees do!
Decision Tree 101 ๐งต
This is exactly what Decision Trees do!
Decision Tree 101 ๐งต
Let's see an example:
You want to decide if you should read a book or not.
The first question is:
"Am I interested in the topic?"
- If the answer is yes, then we follow the process
- If the answer is no, you move on from the book
1/5
You want to decide if you should read a book or not.
The first question is:
"Am I interested in the topic?"
- If the answer is yes, then we follow the process
- If the answer is no, you move on from the book
1/5
If the answer was yes, you can ask
"Are the ratings good on the book?"
- If the ratings are good, you read the book
- If the ratings are bad, you don't read the book.
2/5
"Are the ratings good on the book?"
- If the ratings are good, you read the book
- If the ratings are bad, you don't read the book.
2/5
Now let's see each part of the tree:
1๏ธโฃ At the top, there is the 'root node' or 'root'
It is our first question
2๏ธโฃ 'Branch' - The "Yes" and "No" answers to the question
3/5
1๏ธโฃ At the top, there is the 'root node' or 'root'
It is our first question
2๏ธโฃ 'Branch' - The "Yes" and "No" answers to the question
3/5
3๏ธโฃ Every question gives a 'decision node'.
A decision node has two branches "coming out of it"
4๏ธโฃ The nodes with no branches are 'leaves' or 'leaf nodes'
4/5
A decision node has two branches "coming out of it"
4๏ธโฃ The nodes with no branches are 'leaves' or 'leaf nodes'
4/5
The goal of a Decision Tree algorithm is to find the best questions using the input data.
Decision trees resemble human reasoning.
DTs can be classification & regression models as well.
So they can predict classes (classification) & values (regression).
5/5
Decision trees resemble human reasoning.
DTs can be classification & regression models as well.
So they can predict classes (classification) & values (regression).
5/5
Tomorrow I will share information about the process and how the algorithm searches for questions.
If you don't want to miss it, follow @levikul09
Like/Retweet the first tweet below for support.
Thanks ๐
If you don't want to miss it, follow @levikul09
Like/Retweet the first tweet below for support.
Thanks ๐
Loading suggestions...