1: \begin{abstract}

2: We propose an online tensor subspace tracking algorithm based on the CP decomposition exploiting the recursive least squares (RLS), dubbed OnLine Low-rank Subspace tracking  by TEnsor CP Decomposition (OLSTEC). Numerical evaluations show that the proposed OLSTEC algorithm gives faster convergence per iteration comparing with the state-of-the-art online algorithms. 

3: \end{abstract}

4: