Definition:
A supervised learning algorithm used for classification problems, predicting the probability of a categorical outcome.
Key Points:
Predicts 0 or 1 (binary) or multiple classes
Uses sigmoid function: σ(z)=1+e−z1
Threshold (usually 0.5) decides the class
Applications:
Spam detection
Credit approval
Disease diagnosis
Advantages:
Simple, interpretable
Outputs probabilities
Limitations:
Not suitable for complex/non-linear relationships
Take quizzes related to this topic and see where you stand!
Start Quiz Now