Decision Tree Regression

What is decision tree regression?

The entire data set (parent node) is divided into two data sets (child nodes) based on the reference value. Repeat the above steps for each child node to create child nodes until only one class of data exists in the child node.

There are example images below that show step by step how nodes are made.

Step 1

---

The data set is scattered across a graph with axes X₁ and X₂.

Step 2

---

Make a split 1 by reference value(X₁ < 20)

Step 3

---

Make a split 2 by reference value(X₂ < 170)

Step 4

---

Make a split 3 by reference value(X₂ < 170)

Step 5

---

Make a split 4 by reference value(X₁ < 40)

Example

Code

---

from sklearn.tree import DecisionTreeRegressor
regressor = DecisionTreeRegressor(random_state=0)
regressor.fit(X, y)
X_pred = regressor.predict(X)

Result

---

You can see my implementation on my GitHub