fips.methods.analytic.AnalyticOptions#

class fips.methods.analytic.AnalyticOptions(gamma=1.0)[source]#

The options of the "analytic" method.

Parameters:

gamma (float) – The regularization factor \(\gamma\), which multiplies the observation term of the cost: \(J(x) = (x - x_0)^T S_0^{-1} (x - x_0) + \gamma (z - H x - c)^T S_z^{-1} (z - H x - c)\). It is the same as dividing \(S_z\) by \(\gamma\): above 1 the estimate fits the observations more closely, below 1 it stays closer to the prior. It must be positive and finite.

Raises:

ValueError – If gamma is not a positive, finite number.

Attributes

gamma

the model-data mismatch is divided by it.