Regression Analysis

Regression is a way of estimating an outcome (called dependent variables) given its relationship with predictors (called independent variables).

There are two types of regression:

• Simple regression: when a single independent variable is used to predict the dependent variable
• Multiple regression: when mutiple independent variables are used to predict the dependent variable

Furthermore, regressions can be linear or non-linear regression depending on the relationship between the dependent and independent variables.

Linear Regression

The goal of Linear Regression is to ouput an equation of the form: $y={\theta }_0+{\theta }_1{x}_1+{\theta }_2{x}_2+...+{\theta }_n{x}_n$ optimal for a certain quality metric. Building a Linear Regression Model.

The $\theta$s are the coefficients and x is the input, the independent variable. $y$ is the predicted/estimated value.

Pros

• Very fast
• No parameter tuning necessary
• Easy to understand

Cons

• Limited scope

math