1: \begin{abstract}

2: For the class of stationary Gaussian long memory processes, we study

3: some properties of the least-squares predictor of $X_{n+1}$ based on

4: $(X_n, \ldots, X_1)$. The predictor is obtained by projecting $X_{n+1}$

5: onto the finite past and the coefficients of the predictor are estimated

6: on the same realisation. First we prove moment bounds for the inverse

7: of the empirical covariance matrix. Then we deduce an asymptotic

8: expression of the mean-squared error. In particular we give a relation

9: between the number of terms used to estimate the coefficients and the

10: number of past terms used for prediction, which ensures the

11: $L^2$-sense convergence of the predictor. Finally we prove a

12: central limit theorem when our predictor converges to the best linear predictor based on

13: all the past.

14: 

15: \end{abstract}

16: