Interval recursive least-squares filtering with applications to video target tracking, Optical Engineering | DeepDyve

逐次最小二乗法（RLS: Recursive Least Squares）ーパラメータ同定に用いられる時系列データの回帰手法！ー

Mohammad Javad Azadi on Instagram: "The RLS algorithm is almost one half of the Actor/Critic reinforcement learning agent. It can track nonstationary desired states and adapt to the new data vector with infinite memory in this application. It’s also called the growing window RLS algorithm as opposed to the sliding window variant. The sliding window RLS algorithm has the advantage that it can forget outliers. Its disadvantage is that it doubles the computation. Unicycle implemented the growing wi