Intro to Logistic Regression
- Logistic Regression is a statistical and Machine Learning algorithm to classify data in the dataset.
- it is similar to linear regression, but it takes categorical (discrete) target fields instead the numeric continuous values.
- The classification can b binary (i.e. yes, no), or multi-class.
- Logistic Regression gives the probability of a given class
- The independent variables (features, x) should be continuous (i.e 0.0 .. 1.0)
- The dependent variable (the label, y) should be categorical (i.e. TRUE, FALSE, MAYBE)
What are the applications of Logistic Regression?
- a chance of mortality
- a likelihood of propensity to purchase a product
- a probability for the failure of a process or a product
- a likelihood of default on a loan
When should I use Logistic Regression?
- If the data is binary (i.e. 0/1, Yes/No, True, False)
- If you need probabilistic results
- When you need a decision boundary (linear, polynomial, or more complex hyperplane)
- if you need to understand the impact of the features
