File size: 643 Bytes
b89091f 7558828 b89091f 7558828 b89091f 7558828 b89091f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 |
---
license: cc-by-4.0
---
These are Quantile Random Forest Regression models trained to predict the kpoints-density with different confidence levels:
1. QRF95.pkl predicts (0.05, 0.5, 0.95) quantiles
2. QRF90.pkl predicts (0.1, 0.5, 0.9) quantiles
3. QRF85.pkl predicts (0.15, 0.5, 0.85) quantiles
The performance of models measured for the 0.5 quantile is:
MAE: 0.064,
MAPE: 0.179,
MSE: 0.0098,
R2_score: 0.694,
Spearman_corr: 0.862,
Kendall_corr: 0.682
Models are trained on the dataset generated for this work.
Associated GitHub repositories:
https://github.com/stfc/goldilocks
https://github.com/stfc/goldilocks_kpoints
|