Segway 2.0 Application Note Datasets

Learned parameters and resulting segmentation corresponding to the analyses shown in the Segway 2.0 application note.

Directory structure:

GMM (datasets corresponding to the mixture of Gaussians analysis)

• 1-component
  • traindir/
    • log/ (training log likelihood progression)
    • params/ (learned parameters)
  • identifydir/
    • segway.bed.gz (segmentation)
• 3-component
  • traindir/
    • log/ (training log likelihood progression)
    • params/ (learned parameters)
  • identifydir/
    • segway.bed.gz (segmentation)

minibatch-fixed (datasets corresponding to the minibatch learning analysis)

• fixed/
  • traindir/
    • log/ (training and validation log likelihood progression)
    • params/ (learned parameters)
• minibatch/
  • traindir/
    • log/ (training and validation log likelihood progression)
    • params/ (learned parameters)

TSS_prediction (datasets corresponding to the TSS prediction analysis) (where k=component number=1-5, n=random start number=1-10)

• outputs_[date]_k/
  • traindir/
    • log/ (training and validation log likelihood progression)
    • params/ (learned parameters)
  • identifydir_n/
    • segway.bed.gz (segmentation)