Filters.Kalman1dval create : signal_to_noise_ratio:float -> warmup:int -> tval update : t -> float -> floatUpdate with a new observation; returns the posterior estimate of the state. During warmup Q/R are still being calibrated; the function still returns a sensible smoothed value but ready is false.
val value : t -> floatval variance : t -> floatval ready : t -> boolval recalibrate : t -> unitForce a Q/R recalibration from accumulated residuals — call on regime transitions.