QoM: An effective querying method for time series database

A very promising idea for fast searching in time series databases is to map the time series into a representative space. In this paper, we propose an effective querying algorithm QoM (Querying on Motif) based on time series motifs. First, we map the time series into representative motifs by motif discovery algorithm. Second, we look for a time series data based on these motifs by QoM. The QoM can effectively locate the position of time series by comparing the distance between the time series and central time series of each motif. Meanwhile, in the process of searching time series, we also can further understand the specific time series by the motif which represents a characteristic pattern in the time series database. Furthermore, experiments show that our querying method on databases from diverse domains is feasible and improves efficiency of time series data querying