What is classification?

Classification is the task of assigning data to one of predefined classes.

What is classification?

Classification predicts which of a set of predefined classes an example belongs to: is the e-mail spam, is the tumour benign, which digit is in the photo. The output is not a number but a label.

from sklearn.linear_model import LogisticRegression
X = [[1], [2], [3], [7], [8], [9]]
y = ["düşük", "düşük", "düşük", "yüksek", "yüksek", "yüksek"]
m = LogisticRegression().fit(X, y)
print(m.predict([[2.5], [7.5]]))   # ['düşük' 'yüksek']

Labels can be text; scikit-learn keeps the classes sorted in the classes_ attribute.

Logistic regression

Despite its name, logistic regression is a classifier. It squeezes a linear score between 0 and 1 with the sigmoid function and interprets the result as a probability:

def sigmoid(z):
    return 1 / (1 + np.exp(-z))
print(sigmoid(0))    # 0.5   decision boundary
print(sigmoid(4))    # 0.98
print(sigmoid(-4))   # 0.018

If the probability is above 0.5 the positive class is chosen. It is a fast, interpretable and strong baseline model.

Exercise Sigmoid

Write the sigmoid(z) function with np.exp. Print the results for np.array([-2, 0, 2]) with np.round(..., 3).

Show example solution

import numpy as np

def sigmoid(z):
    return 1 / (1 + np.exp(-z))

print(np.round(sigmoid(np.array([-2, 0, 2])), 3))

Test yourself

Which is a classification output?
- Sales count 340
- "approved" / "rejected"
- Tomorrow's temperature 21.3
- House price 1.2 million
Show answer

"approved" / "rejected" — It is a choice from a limited label set; the others are continuous numbers.
Create the logistic regression classifier.
```
m = ____()
```
Show answer
```
m = LogisticRegression()
```
Match the concepts.

label

Class name
sigmoid

Squeezes into 0-1
0.5

Default decision threshold
classes_

Sorted classes

Next lesson: Binary classification

We Are Enhancing Your Experience

Cookie Preferences

Strictly Necessary Cookies

Analytical Cookies