经典线性回归模型的假定有哪些

1. linear specification
2. regressors X'X is full rank
3. X is non-stochastic
4. innovation is zero mean, homoskedastic and no serrial corrleation
5. innovation follows normal distribution with zero mean, homoskedasticity and no serrial correlation

1,2,3,4 ==》estimators are BLUE (best linear unbiased estimators)
1,2,3,4,5 ==>estimators are BUE
Also, hypothesis testing is based on 1,2,3,4,5.