Study on the Similarity Query Based on LCSS over Data Stream Window

Abstract

Aiming at the problem that the NAIVE algorithm which is taken to handle the similarity query based on LCSS over data stream window (SQLSW) cannot get query results until calculations on all elements in the full dynamic programming matrix are finished, the SQLSW query processing algorithm based on Possible Solution domain optimization strategy (SQLSW-PS) is proposed. It defines possible solution (PS) domain of the dynamic programming matrix about every window. Based on characters of matrix members in the PS domain and the similarity query, it can get query result on the condition that the LCSS similarity function value has not been obtained yet, and reduce lots of computations related to matrix members. It is revealed by extensive experiments that the SQLSW-PS outperforms current algorithms in time, and is effective in handling the SQLSW query.