Conclusion

The complete set of recursions for the general Burg-type adaptive algorithm is reformulated at the end of the paper (appendix B). Though a residual depends on the whole data set, more emphasis is put in the minimization on the residuals near the time of interest, by using an exponential weighting centered on this particular time. However, I expect to achieve good results only if the exponent $\lambda$ is chosen close to 1, since the algorithm appears to me as an approximation of the exact algorithm.