Covariance Smoothing Algorithm Implementation and Data

TReK method introduces a tensorized-and-restricted Krylov algorithm for covariance smoothing with large observational sizes. The algorithm achieves an order of magnitude improvement in computational speed and memory cost over existing methods. Authored by Ho Yun and updated in March 2026.

Use Cases

• Benchmarking covariance smoothing speed against other Krylov subspace methods using provided timing data.
• Analyzing convergence behavior of the TReK algorithm from iteration logs.
• Testing algorithm flexibility with different forward and projection tensor configurations described in the documentation.

Strengths

• Algorithm demonstrates an order of magnitude improvement in speed and memory cost.
• Includes supporting documentation and code files totaling 6.9 MB.

Limitations

• Dataset scale is small at 6.9 MB, limiting the size of included example data.
• Specific row counts and column structures for any example data are unknown.

Provenance

Source

figshare

Collection Method

Algorithm publication and associated implementation files.

Freshness

Last updated March 2026.