1. The regression model is linear in parameters.
2. The mean of residuals is zero.
3. Homoscedasticity of residuals or equal variance.
4. No autocorrelation of residuals.
5. The X variables and residuals are uncorrelated.
6. The variability in X values is positive.
7. The regression model is correctly specified.
8. No perfect multicollinearity.
9. The number of observations, n, is far greater than the number of parameters to be estimated.
9. The variability in X values is positive
10. Normality of residuals