Logistic regression algorithm--solving classification problems

Classification:

Binary classification:

Multi-Category classification: (as multiple two-yuan classification)

Logistic regression model:

Cost function:

Parameter--Gradient descent algorithm

The updates in logistic regression are in the same form as linear regression, but they are different: h (x) is different

In linear regression: H (x) is a linear function

In logistic regression: H (x) is an S-type function

Note: When the feature range gap is large, feature scaling is required.