ANCOVA - is used when our Y variable is a continuous variable
And in the independent(x) variable, we should have at least one Continuous variable and at least one categorical variable.
Linear Regression - is used when our Y variable and X variable both are numeric. Based on the X value the model will try to predict the Y (target).
Multiple Regression - Multiple Regression is an extension of Linear Regression. When we want to predict the Y variable, that depends on multiple X variables then multiple Regression is used.
Logistic Regression -is used when the output is categorical and input could be anything. Common uses are where the dependent value is binary ie the output can be only two things. e:g male:female, true:false or counts.
Logistic Regression is a classification algorithm. So why it is named a Regression?
- As the primary decision made by this algorithm is continuous (regression) and then a final tuning (conversion - mostly using a sigmoid function) convert this continuous output to a categorical output. That's why although it is named as regression, it is a classification algorithm.
0 Comments