Low rank tensor approximation techniques for up- and downdating of online time series clustering KU Leuven
Time series clustering is used to analyse datasets generated by sensors, IoT devices, digital networks, etc. It allows to discover structure and patterns by grouping similar behavior for tasks as explanations, forecasting, and anomaly detection. Current methods struggle to deal with present datasets and needs by users. First, datasets have grown to such scale that clustering requires millions to trillions of comparisons, which one is unable ...