If b = 0, then the prediction of y is always equal to E(Y);

Errors should be distributed by the normal distribution with a mean equal to zero, and variance σ²;

Variable x is called the dependent variable;

The linear regression model does not show that between variables exists linear dependency.