Applied Mathematics Seminar: Harnessing the Power of Time Series Data

Zoom ID: 945 1753 3441
Passcode: USUAMS

Speaker: Soukaina Filali Boubrahimi, Assistant Professor of computer science at USU.

abstract:
Time series data are ubiquitous in daily life. From our heart rates measurements, credit cards transactions, to daily bitcoins pricing, all discrete-valued data that have a time component are considered time series data. The limitation of having huge datasets is not at the storage level, as Moore’s Law predicted it well, but in the ability to extract useful information from it. In this talk, I will expose some ideas on how to harness the power of time series data through some applications such as solar events predictions and spatiotemporal shape interpolation. I will also discuss our recent model, 1DCNN shapelet miner, a novel interpretable motif discovery model inspired from the idea of mining recurring patterns in DNA that are presumed to have a biological function. Finally, I will provide a discussion on some promising directions on multivariate time series classification using shapelets mining algorithms.