scikits.statsmodels.sandbox.tsa.ARIMA

class scikits.statsmodels.sandbox.tsa.ARIMA

currently ARMA only, no differencing used - no I

parameterized as
rhoy(L) y_t = rhoe(L) eta_t

A instance of this class preserves state, so new class instances should be created for different examples

Methods

errfn([rho, p, x]) duplicate -> remove one
fit(x, p, q[, rhoy0, rhoe0]) estimate lag coefficients of ARMA orocess by least squares
forecast([ar, ma, nperiod])
generate_sample(ar, ma, nsample[, std])
predicted([rhoy, rhoe]) past predicted values of time series

Previous topic

scikits.statsmodels.sandbox.tsa.ccovf

Next topic

scikits.statsmodels.sandbox.tsa.ARIMA.errfn

This Page