BTS: a Fast Approach for Similarity Search in Sequences

In order to improve efficiency of similarity search and simplify process of index building, a new scheme was proposed: BTS. BTS introduced bit trend representative into the feature space along with significance segmentation and similarity model. Bit trend representative was a fast similarity model, in which trend match was taken to exclude many false alarms through quick bit operations. On the other hand bit trend representative could act as an efficient index and there was no need for an independent index building process. Experiments showed that BTS could identify kinds of transformations and ran much faster than other shape-based approaches. BTS can improve search efficiency significantly