3.9. Extension: Predicting With a Classification Tree

Consider the following test data:

Weight (kg)	Tail Length (cm)	Ear Length (cm)	Class
4	25	4	cat
3	3	7	rabbit
3.5	7	3	cat
27	22	7	dog

We can use .predict(x) to use our decision tree classifier to predict the classes of our test data. We just need to remember to provide the test inputs as a 2D array where the columns correspond to each input variable and the rows correspond to each sample.

import pandas as pd
import numpy as np
from sklearn.tree import DecisionTreeClassifier

data = pd.read_csv("animals.csv")
x = data[["Weight (kg)", "Tail Length (cm)", "Ear Length (cm)"]].to_numpy()
y = data["Class"].to_numpy()

tree = DecisionTreeClassifier(max_depth=3)
tree.fit(x, y)

# weight, tail, ear
x_test = np.array([[4, 25, 4], [3, 3, 7], [3.5, 4.5, 3], [27, 22, 7]])

prediction = tree.predict(x_test)
print(prediction)

Output

[0 1 0 1]

We can see the prediction is

[0, 1, 0, 1]

This means:

• the first sample has been classified as class 0: cat

• the second sample has been classified as class 1: dog

• the third sample has been classified as class 0: cat

• the fourth sample has been classified as class 1: dog

We can verify these predictions with the models ourselves as well using the diagram.

Code Challenge: Extension: Predicting with a Classification Tree

Now lets use the classification tree we just built on our astronomy data astronomy.csv to classify the astronomical objects in our test data.

Name	Mass (Earths)	Density (g/cm:math:^3)
Kepler 1649	1.2	5.54
Eris	0.0027	2.43
Io	0.015	3.53
Miranda	0.00001	1.15
TrES 4b	248	0.16

Instructions

1. Copy and paste in your code from Extension: Interpreting The Output Graph, just up to where you fit the classification tree

2. Create a numpy array containing the astronomy data shown above

3. Use .predict to predict the class for each object

4. Print the predictions

Your output should look like this:

[X X X X X]

Solution

Solution is locked