[Numpy-discussion] poly class question

[josef.pktd at gmail.com]

Is there a way in numpy (or scipy) to get an infinite expansion for
the inverse of a polynomial (for a finite number of terms)

np.poly1d([ -0.8, 1])**(-1)

application for example the MA representation of an AR(1)


and fractional powers

np.poly1d([ -1, 1])**0.5

this is useful for fractionally integrated time series, e.g. ARFIMA

Until now I did this directly or using scipy.signal, but I thought
maybe the polynomial class would handle some of it, both examples
raise exceptions.

Josef