The linear regression model has several assumptions, including:
Homoscedasticity
Data that exhibits homoscedasticity is appropriate for linear regression. If homoscedasticity is violated, the dataset may need to be changed or a different model may need to be selected.
Autocorrelation
The dataset should not be autocorrelated, which means that residuals or error terms are independent of each other.
Errors are normally distributed
The residuals should be distributed around zero for the entire range of predicted values. If the residuals are evenly scattered, the model may perform well.
Independence
The independent variables in the model are not correlated or related to one another. This assumption is necessary for linear regression to be valid.
Independence of errors
Residual errors are independent of each other.
Lack of perfect multicollinearity
This refers to a situation where a number of independent variables in a multiple regression model are closely correlated to one another.
Linear relationship
There is a linear relationship between the predictors (x) and the outcome (y). Variable transformation may be necessary to fulfill this assumption.
Omitted variable bias
The error term is uncorrelated with the regressors. The presence of omitted variable bias violates this assumption and causes estimators to be bias and inconsistent.
