Regression Analysis: Cause that effect
Linear regression: a linear approximation of a causal relationship between 2 or more variables
Linear regression is the pure form of algorithm which correlates between two variables in the data set. The input and output sets examined to show a relationship. It also shows how the change in one variable can affect the other variable. It is represented by plotting a line on the graph. The algorithm is popular because it is easy to explain, transparent, and requires no tuning. Companies use this algorithm to forecast sales and risk assessment to take long term business decisions.
Process:
• Get sample data
• Design a model that works for that sample
• Make predictions for whole new variable
Simple linear regression model: function: y = b0 + b1x + e
Where, b0 is constant or bias, b1 is coefficient, x is predictor or independent variable,
y is predicted or dependent variable, e is epsilon or error
Simple linear regression equation: y^ = b0 + b1x1
y^ is predicted or estimated value
Correlation: measure degree of relational in 2 variable, in graph it is single point
Regression: how one variable affect other, one way, in graph it is line
Regression line: best fitting line through data points
Residual: estimator of error
