Abstract:
Time-series shapelets, discriminative subsequences, are effective for time series classification (tsc). Tsc accuracy depends on shapelet quality. However, major research has focused on accurate models from some shapelet candidates.
Existing studies use simple methods like enumerating subsequences of fixed lengths or randomly selecting shapelet candidates. Building the candidate model takes most of the computation. This paper introduces bspcover, an efficient shapelet discovery method that finds high-quality shapelet candidates for model building.
Specifically, bspcover generates abundant candidates via Symbolic Aggregate approXimation with sliding window, then prunes identical and highly similar candidates via Bloom filters and similarity matching.
Next, a pp-Cover algorithm efficiently finds discriminative shapelet candidates that maximally represent each time-series class.
Finally, a classification model can be created using any shapelet learning method. We tested popular time-series datasets and cutting-edge methods. bspcover speeds up state-of-the-art methods by more than 70 times and often outperforms them in accuracy.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
