Table 1 Optimal hyper-parameters for each case
Case | ABIC | \({\alpha }^{2}\) | \({\sigma }_{1}^{2}\) | \({\eta }^{2}\)(strain/tilt) |
|---|---|---|---|---|
Synthetic data | − 445.38 | 4.16 × 10–15 | 5.08 × 10–20 | 0.92 |
Example 1: 14–19 April 2015 | 2.45 | 1.26 × 10–10 | 1.80 × 10–15 | 0.38 |
Based on stain data alone | 229.32 | 3.40 × 10–12 | 1.01 × 10–16 | – |
Based on tilt data alone | 420.14 | 8.11 × 10–11 | 2.16 × 10–15 | – |
Example 2: 29 December 2015 to 1 January 2016 | − 336.63 | 4.35 × 10–11 | 6.23 × 10–17 | 0.10 |
Example 2: 1–5 January 2016 | − 284.53 | 1.64 × 10–11 | 2.20 × 10–16 | 0.25 |
Example 2: 5–7 January 2016 | − 320.50 | 4.81 × 10–11 | 1.10 × 10–16 | 0.91 |
Example 3: 23–27 October 2016 | 547.83 | 6.85 × 10–11 | 8.01 × 10–16 | 0.46 |
Based on stain data alone*1 | 780.51 | 8.74 × 10–10 | 1.33 × 10–15 | – |
Based on tilt data alone*1 | 702.08 | 1.82 × 10–11 | 5.73 × 10–16 | – |