Research map / Time series / Causal and generative time series
LSAR: Efficient Leverage Score Sampling Algorithm for the Analysis of Big Time Series Data
Ali Eshragh, Fred Roosta, Asef Nazari, Michael Mahoney
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use fully-approximate leverage scores instead of SRHT | Leverage score approximation on synthetic AR(20), AR(100), and AR(200) · Synthetic AR(20), AR(100), AR(200) | MPRE: around 3 -> around 0.1670 | Improved |
| Sample 0.1% of rows using LSAR rather than using the full data matrix | LSAR on synthetic AR(20), AR(100), and AR(200) · Synthetic AR(20), AR(100), AR(200) | PACF order estimation | No clear effect |
| Use leverage score sampling instead of uniform sampling | LSAR on gas sensor data · Gas sensors data (R8) | PACF order estimation | Improved |
| Use leverage score sampling instead of uniform sampling | AR(16) parameter estimation on gas sensor data · Gas sensors data (R8), AR(16) | relative error of parameter estimates | Improved |
| Use leverage score sampling instead of uniform sampling | AR(16) parameter estimation on gas sensor data · Gas sensors data (R8), AR(16) | ratio of residual norms | Improved |
| Increase the sample size | Fully-approximate leverage score estimation for synthetic AR(100) · Synthetic AR(100) | MPRE | Improved |
| Increase the data size at a fixed sample-to-data ratio | Fully-approximate leverage score estimation for synthetic AR(100) · Synthetic AR(100) | MPRE | Improved |
Closest papers by meaning
- Importance Sampling for Nonlinear Models (ICML 2025 · similarity 0.79)
- A Set-Sequence Model for Time Series (NeurIPS 2026 · similarity 0.79)
- Real-Time Selection Under General Constraints via Predictive Inference (NeurIPS 2024 · similarity 0.79)
- Sampling by averaging: A multiscale approach to score estimation (NeurIPS 2025 · similarity 0.79)
- Turnstile $\ell_p$ leverage score sampling with applications (ICML 2024 · similarity 0.79)
- ChronoEpilogi: Scalable Time Series Selection with Multiple Solutions (NeurIPS 2024 · similarity 0.78)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
