Regression analysis
Regression analysis
- Regression like , relation is a study of a relationship between variables
- Simple regulation : single explanatory variable
- Multiple regression : includes any number of explanatory variables
- Dependent variables : The single variable being explained/ predicted by the regression model
- Independent variable: Explanatory variables used to predict the dependent variable
- Residuals :Portions of dependent variables is not explained by the model
- Correlation tests or closeness and direction of relationship between the two phenomena . Regression gives the nature and extent of this relation, so that prediction can be made about dependent variable, for any value of independent variable.
- Clearly designates one variable as cause and the other effect , unlike the correlation
- Linear regression :straight line relationship, Y=mx+c
- In least square method of regulation ‘m’ and ‘b’ are calculated by equating the differencal of Loss Function (Yi-Pi) to zero
- Formula:-
= | dependent variable | |
= | function | |
= | independent variable | |
= | unknown parameters | |
= | error terms |
- Non linear regression : implies third relationships. e.g Logarithmic relationships
- Cofficient of determination(R^2) : Sum of sq explained by a regression /total number sum of Squares. It lies between 0 and 1 .if R^2=1 it means 100% of variation in Y is explained by variation in X
Graph :-
Application:-
- Spectrophotometry and similarother and analytical techniques
- Used to model real life situation for which outcomes is uncertain. Modelling the appropriate probability distribution can help to make predictions and inferences.
- Probability of an event varies between 0 and 1